Government Executive - Tech

CMS hopes to quadruple provider suspensions over next 6 months, official says

Christian Robles — Fri, 16 Oct 2026 08:38:00 -0400

SANTA FE, N.M. — The Centers for Medicare and Medicaid Services plans to expand provider suspensions and publish the results of a pilot testing whether artificial intelligence can prevent Medicaid fraud, a senior agency official said Thursday.

The two moves are the latest in the Trump administration’s technology-enabled drive to trim fraud, waste and abuse in federal health care programs. The push has helped yield $41.9 billion in savings for fiscal year 2025, or a 59% increase over the previous year, according to a slideshow CMS Deputy Administrator Kimberly Brandt presented at the ADAPT conference in Santa Fe, New Mexico.

In April 2026, CMS launched a “Medicaid War Room” to curtail fraud, waste and abuse. The CMS arm has stopped $342 million in improper payments as of Sept. 2026 and sent 124 notices of intent to exclude providers, according to the slideshow.

“My hope is that suspension number will quadruple over the course of the next six months,” following new exclusion authority granted by the Health and Human Services Department earlier this week, Brandt said.

With that power, CMS can kick providers out of Medicare, Medicaid, TRICARE and other federal health care programs. Brandt previously said she urged congressional offices to give CMS the exclusion authority.

Beyond CMS’ push to exclude certain providers, Brandt noted that the agency is piloting AI and machine learning to detect Medicaid fraud.

The pilots use AI to flag suspicious billing patterns — such as upcoding, phantom services and geographic anomalies — in near real-time before Medicaid payments are made, according to the slideshow. The pilots are also testing a shared platform where CMS, Medicaid Fraud Control Units, program integrity contractors, states and law enforcement work on high-risk cases together.

“If proven out, the goal is to scale this into a national, standardized, continuously improving platform that stops fraudulent Medicaid payments before they’re made,” the slideshow said.

Brandt said CMS will soon publish the results of an Indiana Medicaid fraud pilot the agency conducted with Oracle, and that it is also actively working with five other states on Medicaid pilots. But the agency is also using other technologies to further its mission to combat fraud, waste and abuse.

CMS is utilizing a “Netflix-type algorithm” that compares new providers and suppliers with ones the agency already has a relationship with, Brandt said. The technology allows CMS to put suspicious providers and suppliers on a watchlist for the first six months after they enter a federal health care program. The agency has cut about 67% of those providers and suppliers because they did not follow or comply with program rules, she said.

Earlier this week, the Trump administration announced it cut subsidies for 189,000 people on the Affordable Care Act exchanges after it found those same enrollees had duplicate health care coverage or were dead. The administration said it found some of the duplicate beneficiaries through periodic data matching of Medicaid and the Children’s Health Insurance Program.

Brandt added that the agency is “doing even more” analysis of the Social Security Administration’s Death Master File so as not to issue payments to dead people, though she did not elaborate on what the analysis was.

Even so, CMS’ use of technology to claw back fraud has not been perfect, Brandt acknowledged.

“There are a lot of things about WISeR that are not working,” Brandt said, referring to a controversial pilot of AI-supported prior authorizations of Medicare items and services that have a high risk for fraud.

Hospitals and Democratic lawmakers have criticized the six-year initiative for delaying and denying seniors’ access to care since it launched at the beginning of the year. The House Ways and Means Committee advanced a provision to end the WISeR program in June, while the Senate has rejected efforts to stop it.

“Especially because of the congressional scrutiny on [the WISeR program, CMS Innovation Center staff] are committed to trying to get that right and get it better,” Brandt said.

Among the lessons CMS has learned from the WISeR model are that technology requires continuous oversight, human clinical review is essential and provider feedback is an important monitoring signal, according to the slideshow.

]]>

AI offers Secret Service unique mission enhancements, official says

Edward Graham — Thu, 15 Oct 2026 16:41:16 -0400

Artificial intelligence is providing the U.S. Secret Service with new capabilities to carry out its protective and investigative missions, the agency’s IT and AI lead said on Thursday, with further adoption of the technologies poised to help it respond even better to future risk environments.

Given the Secret Service’s directive to both safeguard top U.S. officials and visiting dignitaries and to conduct criminal investigations, Chief Information Officer and Chief AI Officer Chris Kraft said “we have to look broadly at what's the art of the possible” when it comes to leveraging AI for its operations.

While Secret Service, like other federal agencies, has leveraged AI to streamline mundane back-office operations or other rote tasks — Kraft noted during the SAP Public Sector Summit in Washington, D.C., that “with generative AI, there's huge opportunity to automate things, such as invoice processing or contracts or other capabilities” — its unique mission positions it to utilize some of the more cutting-edge AI capabilities.

The service is trying to navigate an unprecedented threat environment with an understaffed workforce. During an interview with 60 Minutes on Sunday, Secret Service Director Sean Curran said threats against the agency’s full-time protectees are up more than 40% since last year, with around 10,000 threats made on those individuals since January.

“I believe that artificial intelligence can help us as we investigate those threats, monitor those threats [and] bring together correlations between seemingly disparate information,” Kraft said, referencing Curran’s interview. “I also think when you look at facial recognition and biometrics, [you see] the things that you can do — and at scale — with artificial intelligence.”

Kraft said, for instance, that an object detection tool could help the agency quickly identify how many individuals are in one space “at a speed that a human is not able to do that.”

The Secret Service currently lists three AI use cases in its most recent public inventory, with two of those uses — a model to support criminal investigations through facial recognition matches and another that is used to decipher license plates from poor-quality images or videos — currently deployed by the agency. Kraft said the agency’s AI adoption is meant to empower Secret Service’s human agents to better carry out their duties.

In an interview with Nextgov/FCW earlier this year, Kraft also said the Secret Service is working to embed AI specialists across its operations to help drive further adoption of AI tools.

Kraft highlighted that initiative on Thursday and said developing an internal AI team will allow experts to “go and work with different parts of the organization and accelerate, where it makes sense, our adoption of the technology.”

]]>

OPM revamps online pay tables, spearheads AI adoption across federal workforce

Jory Heckman — Wed, 14 Oct 2026 17:18:32 -0400

The federal government’s human resources agency is revamping one of the most heavily trafficked tools on its website: annual pay tables for federal employees.

The Office of Personnel Management is planning to launch a new version of the online tables that agencies and employees use to look up federal pay rates each year.

Adam Starr, the agency's chief information officer, said on Wednesday that the agency’s pay rates website gets about 14 million visits each year, but hasn’t changed much since 2011.

“I always like to remember that 'OPM' really stands for ‘other peoples’ money.’ And in this case, it is your money. People come to OPM pay tables, asking us to show them the money, and for years these have been legacy, clunky sites,” Starr said Wednesday at the SAP Public Sector Summit sponsored by GovExec. “It’s not the experience that a federal employee deserves.”

Starr said the revamped OPM pay tables will launch “later this calendar year,” in the first quarter of fiscal 2027.

“We’re launching it in beta. This will be the default, but people can always fall back to the old version if they don’t like this new version. We’ll get feedback from the agencies, from the employees. We can iterate, relaunch, and get this in people’s hands. Because that’s how everyone’s doing it in industry, and government should be no different,” he said.

Starr said OPM tested the pay table redesign with an accessibility expert who uses a screen-reader. He said that individual “had absolutely no issues using it.”

“For too long, I think civil servants have really suffered with outdated technology that doesn’t keep at the frontier,” Starr said, adding that OPM is looking to give employees the “latest and greatest tools.”

OPM’s pay tables are the latest in a series of government website refreshes happening under the Trump administration. The agency last year overhauled its federal workforce data website, and the White House’s Nation Design Studio is on a mission to transform more than 27,000 federal websites.

These efforts culminated recently in the launch of America.gov, a site that allows users to ask a chatbot questions about everything from qualifying for Medicare to finding a job to renewing a passport.

Starr described the OPM pay-table refresh as part of a broader plan to “make every part of the federal experience tech-first,” and to get federal employees “the best support that you need in every stage of your career.”

“This is kind of just an example of many of the modernizations that we’re trying to do across OPM, from consolidating 119 different HR IT systems to answering every phone call in our retirement services contact center,” Starr said.

OPM is also spearheading some of the Trump administration’s latest efforts to foster greater adoption of AI tools across the federal workforce. President Donald Trump signed an executive order last month rebranding AI across the federal government as “super intelligence.”

Starr said OPM has launched a consulting program called the SuperIntelligence Governance and Adoption Program (SI GAP), focused on helping other agencies adopt AI tools.

“Every SI transformation industry has HR at the table, and OPM is HR for the government,” Starr said. “Really, it comes down to being able to take measured risk of trusting your

employees with these tools, giving them the opportunity to experiment and reinvent their work — because just using it as a better chatbot or a better Google search is not going to reinvent the way that we have the enterprise work.”

Starr said that agencies – including OPM – have historically been too risk-averse with AI, and haven’t yet seized the “opportunity to reinvent their work” with these tools.

“I imagine a world where every employee is more empowered to do the impactful work that they need to do — that a lot of the tedium of the day can be automated,” he said.

Starr said OPM recently launched an AI agent to help process phone calls for the retirement services team.

“We showed it to the entire agency. People were aghast at how human it sounded, how helpful it was,” Starr said. “Literally today we released a new feature, and we released a new feature on Friday last week. And every day this is getting better.”

The federal retirement backlog under the second Trump administration has drawn scrutiny from House Democrats. The backlog sat at 55,681 pending cases on March 31, down from a peak of 65,237 the previous month. The backlog fell to 8,814 pending cases in September.

“We've historically had a large backlog of retirement claims, and it's not because the employees don't want to do good, but there's been processes that are highly optimized on paper that just can't improve until they're digital. They're now digital, and so how can we now optimize those and lead to faster retirements? How can we onboard people quicker?”

OPM in a recent memo directed agencies to use AI to speed up the hiring process for new employees employees. OPM Director Scott Kupor has also been tapped to serve as vice chairman of the administration’s interagency Super Intelligence Task Force.

Starr said OPM understands the importance of keeping guardrails around the data it holds. OPM in 2015 was the target of a massive data breach, exposing about 22 million federal records.

“We have a high risk tolerance for delighting our customers and a low-risk tolerance for security breaches,” he said.

OPM, he added, has three basic rules for AI use: employees are responsible for the work they produce with AI, can use only approved tools, and must stay signed in with official accounts.

“It’s on the CIO team to make sure this is secure, and just constantly thinking about what are the risks and how do we mitigate it. There are concerns about access to the internet via these tools, so we put in a cached version, so nothing hits the internet,” Starr said. “[We’re] just constantly looking at what are the concerns, knocking that down, and iterating and improving.”

]]>

VA rolls out CLEAR login option and in-person identity verification pilot

Edward Graham — Thu, 08 Oct 2026 18:20:00 -0400

The Department of Veterans Affairs announced on Wednesday that it has rolled out a new identity verification login option from CLEAR and established a pilot program at six medical facilities to help veterans create online VA accounts in person.

The new services are billed as part of a broader digital modernization effort being spearheaded by GovCIO to enhance VA’s online services and integrate the CLEAR1 security identity platform onto the VA.gov website.

Veterans now have the opportunity to voluntarily create CLEAR accounts to access their VA services using facial recognition. The agency’s platform also offers other log-in services that use biometrics for identity verification, like Login.gov — the government’s single sign-on service for federal and state platforms — and ID.me — another private sector offering that can be used to verify users of federal platforms.

Veterans who create CLEAR accounts can also use the service to sign into Medicare.gov through a similar partnership announced by the Centers for Medicare and Medicaid Services earlier this year.

The CLEAR accounts used to access VA.gov are different from those operated by the company that allow users to access CLEAR-verified airport security lines.

“Veterans want better customer service and convenience, and under President [Donald] Trump, VA is delivering,” VA Secretary Doug Collins said in a statement. “This technology is another way we’re changing VA to put Veterans first and make it easier for those who have worn the uniform to get the care and benefits they’ve earned.”

In addition to the new login option, VA is deploying CLEAR pods at selected medical sites to help veterans create identity verification accounts while at the facilities. Veterans who visit these biometric verification kiosks with the proper identifications can have their photos taken and then used to create CLEAR accounts.

The on-site identity verification pilot is now active at VA facilities in Atlanta, Georgia; Brooklyn, New York; Orlando, Florida; Palo Alto, California; San Juan, Puerto Rico; and Washington, D.C.

During a Wednesday interview on CBS Mornings, CLEAR CEO Caryn Seidman Becker said veterans can go to VA.gov and verify their identities with “one selfie, and you’re in.” She said the CLEAR pods’ pilot program will also allow the firm “to meet veterans where they are and help them to enroll there.”

]]>

GSA still has problems with location data for federal properties, watchdog finds

Christian Robles — Tue, 06 Oct 2026 17:52:00 -0400

The General Services Administration is still not meeting the Geospatial Data Act of 2018’s requirement to maintain geospatial data, an Office of Inspector General report published Monday found. Previous 2024 and 2022 OIG audits also found that GSA was not fully compliant with the federal law.

GSA has two publicly available geospatial datasets — the Inventory of Owned and Leased Properties Dataset and the Federal Real Property Profile Dataset — that contain inaccurate or missing latitude and longitude coordinates, the OIG found.

The watchdog noted that about 78,624 — or 26% — of properties in the FRPP dataset with coordinates lacked distinct ones and “numerous properties” in that dataset contained erroneous coordinates. The finding echoes a June 2026 Government Accountability Office report that determined about 20% of buildings in the FRPP had coordinates that did not match their reported zip codes.

The OIG found that 634, or about 7%, of the GSA-controlled properties listed in the IOLP-buildings dataset lacked distinct coordinates. Approximately 20% of the IOLP-buildings dataset’s real property assets contained outdated and incorrect congressional representative data.

“GSA should identify cost-effective measures to improve the accuracy and usefulness of the latitude and longitude coordinates in the IOLP and FRPP datasets based on each property’s physical location,” the OIG wrote.

GSA did not maintain its geospatial data inventory for two years until it responded to OIG’s audit, according to the report. The updated inventory showed a roughly 22% increase in geospatial data assets compared to the 2024 inventory.

The OIG found that the agency did not maintain an updated inventory because it does not have a documented process listed in materials like the GEOSPATIAL DATA Standard Operating Procedure — or SOP — and Standards Guide.

“GSA should update its geospatial data SOP to establish a process for maintaining its geospatial data inventory, including defining roles and responsibilities,” the office wrote.

While the Geospatial Data Act requires agencies to search for existing geospatial data before procuring new data, the OIG found that GSA continues to not comply with that requirement.

Following the 2024 audit, GSA agreed to establish a search process, but the agency has not yet implemented it. GSA staff said in April that they were “evaluating resources to determine a timeline for updating the process,” the OIG wrote.

The report said that GSA’s delegation of Geospatial Data Act responsibilities is not effective, leading to some of the recurrent issues. GSA staffing cuts since 2025 have exacerbated those problems, the report added.

GSA did not respond to a request for comment on the OIG report by the time of publication.

GSA Chief Information Officer David Shive did sign a letter agreeing with all five of the OIG’s recommendations for improving geospatial data. The letter says GSA is “working on a plan to address” the recommendations.

Shive agreed to conduct a cost-benefit analysis of feasible corrective action plans for geospatial data accuracy, create a process for correcting IOLP congressional representative name data and establish protocols for GSA-wide compliance with Geospatial Data Act rules.

Shive also agreed to align the GSA’s geospatial strategy plan with a federal 2025-2035 strategic plan and update GSA’s GEOSPATIAL DATA SOP and Standards Guide.

]]>

FBI removes Accenture contractor after missed security patch led to breach

David DiMolfetta — Tue, 06 Oct 2026 16:29:20 -0400

The FBI has severed ties with a contractor working for Accenture amid a massive cybercrime intrusion that may have exposed data on thousands of bureau employees, according to a person with knowledge of the matter.

Prolific cybercrime group ShinyHunters claimed responsibility for the hack last month. The data stolen contained employees’ addresses, phone numbers and information on their spouses, among other categories. It also revealed sensitive data on employees’ intelligence and surveillance roles, as well as private medical information.

Accenture is responsible for software patch management and maintaining custom code at the FBI, said the person, who spoke on the condition of anonymity because the situation is sensitive. Oracle — whose PeopleSoft platform was the initial access point that ShinyHunters claimed to have exploited — provided the security patches that were not integrated, the person added.

FBI cybersecurity chief Brett Leatherman confirmed the removal of an unnamed contractor in a statement to Nextgov/FCW.

“To date, our review has determined that the incident occurred as the result of a security failure of a platform managed by a third-party organization — after a contractor failed to implement a security patch explicitly issued to secure the platform,” Leatherman said. “As such, the FBI has removed the contractor and taken all necessary steps to both mitigate any further risk and protect our workforce.”

The remarks offer the bureau’s clearest explanation yet of how the intrusion occurred, though they do not address why the patch was missed and how the FBI monitored the contractor’s work to ensure systems holding sensitive employee information were protected.

Reuters first reported the removal late Monday. The specific Accenture employee could not be immediately identified.

In a statement, the company said it’s “proud to support the mission of the FBI and will continue to do so.”

The breach follows recent findings from Google’s Mandiant that ShinyHunters had resumed widespread exploitation of a PeopleSoft vulnerability for which Oracle issued a patch in June.

The incident poses serious counterintelligence risks because the exposed records could help foreign intelligence services identify employees working on sensitive investigations and exploit personal information about them.

ShinyHunters has said it would not publish the stolen data, but its statement did not say the records had been deleted. Retired Lt. Gen. Robert Skinner, who previously led the Defense Information Systems Agency, told Nextgov/FCW last week that the group could still sell some or all of the information to foreign intelligence services or other buyers.

Authorities have also moved against suspected members of the group. Dutch police last week announced the arrest of an alleged leader, and Reuters reported Saturday that another suspected member, Saif al-Din Khader, had been detained in Jordan and was helping investigators locate other hackers.

]]>

US spy chief tapped as White House AI czar

David DiMolfetta — Mon, 05 Oct 2026 14:08:13 -0400

President Donald Trump named intelligence director Jay Clayton to serve as the White House AI czar, just two months after being sworn in as the nation’s top spy.

Clayton will head what Trump referred to in a Truth Social post as the “Super Intelligence Force” meant to coordinate the government’s engagement with various organizations, critical infrastructure providers and AI companies. The Trump administration has begun referring to AI as “super intelligence” after the president signed an executive order late last month calling for federal agencies to use the new phrasing.

According to its charter, the task force will create plans to handle AI-enabled threats, streamline breach notifications and improve government response capacities while aiming to avoid overregulation. Its creation follows growing public concerns over AI safety and cybersecurity risks, as well as heated political discussions over data center buildouts that power many of today’s AI systems.

“The risk of not being first is high,” Clayton told the Wall Street Journal in a Saturday exclusive that discussed the task force. “Not being first increases the identified, and unidentified, risks, particularly from our adversaries,” he said. “Being first will better enable us to address those risks on behalf of the American people.”

The task force features three vice chairs: Undersecretary of Defense for Research and Engineering Emil Michael, Office of Personnel Management Director Scott Kupor and Federal Trade Commission Chairman Andrew Ferguson. Its members include Vice President JD Vance, Defense Secretary Pete Hegseth and other senior advisors, alongside external advisors like venture capitalist David Sacks — who previously worked as Trump’s AI and crypto czar early in the administration — and former Secretary of State Condoleezza Rice. It will have 120 days to report on the risks and opportunities of AI and establish a plan for the federal government to tackle it.

The panel’s establishment marks a shift for an administration that has predominantly maintained a hands-off approach toward AI governance. But officials have faced mounting pressure over the past six months to respond as a new generation of sophisticated, autonomous AI models demonstrating advanced cybersecurity capabilities have repeatedly circumvented or escaped established technical safeguards.

Its creation highlights a steady group of voices in the administration and tech industry that believe the time for AI safeguard discussions should begin, though Trump and other officials have largely continued to lean toward a more relaxed approach to regulating the technology.

In public discussions, officials and researchers typically frame AI development as a trade-off between slowing down for safety and accelerating to maintain an edge over foreign adversaries pursuing AI for military and economic gains.

China has especially become a major player in the global AI race. Clayton’s appointment to the task force appears to signal a shift by the administration to cast AI as a geopolitical and national security subject area.

His dual role is a unique move for the White House, placing the current DNI — who oversees all 18 U.S. spy agencies — at the helm of broad domestic technology and economic policy discussions. Clayton brings a background as a corporate lawyer and former SEC chairman, making him a unique pick compared to people with more tech-focused backgrounds that hail from Silicon Valley.

The U.S. intelligence enterprise is poised to benefit from advanced AI adoption. Integrating AI has already transformed how spy agencies digest vast troves of collected data and build intelligence products, among other things.

Sen. Mark Warner, D-Va., the vice chairman of the Senate Intelligence Committee, said that he welcomed the administration’s attention toward the fast-evolving technology but said Clayton’s new title lacks a coherent strategy.

“We still do not know what authority this new role will have, whether it will include developing safety standards or reviewing the most advanced models, or how it fits into the administration’s broader approach to AI,” Warner said in a statement. “That lack of clarity is especially concerning as leading AI companies themselves are confronting increasingly capable systems that can evade safeguards and exploit cybersecurity vulnerabilities.”

“I hope this appointment is an opportunity for the administration to move beyond slogans and toward a serious, coordinated national strategy for AI,” he added.

]]>

Transportation works to ease fears about its new AI air traffic control tool

Edward Graham — Mon, 05 Oct 2026 08:00:00 -0400

The Transportation Department’s request to receive an additional $30 billion from Congress to push its massive air traffic control modernization project across the finish line would include new funding to finish overhauling existing IT systems and deploying new artificial intelligence capabilities.

Reactions to the recent rollout of new AI capabilities at Washington, D.C.-region airports, however, have underscored some of the broader national concerns about relying more on advanced technologies, with officials having to go out of their way to ease worries about the tools.

Transportation and the Federal Aviation Administration, which is overseeing the ambitious effort to modernize management of U.S. airspace, announced on Sept. 21 that they officially rolled out a new AI-powered tool for Ronald Reagan Washington National Airport, Dulles International Airport and Baltimore/Washington International Thurgood Marshall Airport that combines 200 different data streams into one system that can provide better flight traffic management.

FAA said at the time that the Strategic Management of Airspace, Routes and Trajectories — or SMART — tool synthesizes information like flight paths and weather patterns “to provide a comprehensive visualization of where planes are going, how much traffic the system can handle, and where congestion or weather could cause challenges.”

During a Monday press conference announcing the department’s request for $30 billion, Transportation Secretary Duffy cited SMART’s recent deployment and said it is “giving different tools to be predictive in what we not just see at this moment, but what we see in the future, because we want to deliver better for the American people, reduce their delays and reduce their cancelations.”

Duffy added, however, that while new tools like SMART can help enhance air traffic control management, the FAA's continued reliance on outdated systems means that more underlying tech modernization efforts are needed.

“I can operate SMART on that old operating system, but I want a new operating system so I can give new tools to the air traffic controllers,” he said.

Despite Transportation’s excitement about further deployments of the SMART tool, the fact that it is AI-powered has led to some unease from lawmakers and others.

Last month, prior to SMART’s rollout, Politico reported that airline industry officials had been “panicking” about the tool’s integration with other air traffic control systems before the FAA ultimately alleviated their concerns about the new capability.

Rep. Don Beyer, D-Va., who represents the congressional district in which Ronald Reagan Washington National Airport is situated, expressed similar concerns and released a statement the day the SMART tool was set to be deployed that called for its impending rollout to be suspended, saying that “the FAA should not use my constituents as guinea pigs for an unproven AI air traffic control system.”

In an interview with Nextgov/FCW following SMART’s rollout at the D.C.-area airports, Beyer referenced growing global concerns about the capabilities of more advanced AI models and said his office received feedback expressing similar misgivings about the tool’s deployment.

“If it was just the constituents it would be one thing, but then the air traffic controllers were jumping in saying, ‘Hey, we don't know anything about this. We've not been part of the design. We didn't see anything ahead of time. We found that at the same time you did that it’s going to suddenly be made operational,’” he said.

Beyer said he subsequently spoke with FAA Administrator Bryan Bedford to share his concerns and had “a good conversation” with him about the SMART tool, including how it was created.

In a statement to Nextgov/FCW, the National Air Traffic Controllers Association said it “has not been involved in the design, testing or implementation of SMART and therefore cannot speak to its effect on the air traffic control workforce at this time.”

But Beyer said Bedford was reassuring and told him that the tool was designed by air traffic controllers who aren't in the towers right now but are the “30- and 40-year guys who've been doing it all their lives.” He added that Bedford also said SMART is operating at the D.C. region’s Terminal Radar Approach Control facility in Warrenton, Virginia, and not within the towers at Dulles or National airports.

“I think the feedback was you almost need big, busy airports in order to generate enough information to make AI relevant,” Beyer said about SMART first being deployed for the large airports in the Washington, D.C. region, instead of being piloted at smaller airports.

He added that he told Bedford “it would have been good to explain that to the American public,” particularly given the current global discussion around new AI capabilities.

In response to a request for comment, Transportation pointed to a Sept. 22 tweet from Duffy, in which he reposted Beyer’s release expressing concerns about the tool’s deployment and wrote: “Here’s the truth: HUMANS still control the airspace and HUMANS still make the calls. With SMART, we are giving air traffic controllers new technology and better data to make critical decisions.”

In a Fox News clip that was shared with the post, Duffy called SMART “a tool that’s used to better manage traffic” and said air traffic controllers “are the ones who will always manage airspace.”

Recent high-profile incidents where frontier AI models have taken undirected action — including one case in which OpenAI agents escaped from their testing environment and hacked into Hugging Face on their own volition — have exacerbated concerns from tech executives, lawmakers and others about whether guardrails are needed on advanced AI technologies.

President Donald Trump has been vocal about his desire to not overregulate the AI sector to ensure that the U.S. continues to outpace China in development of the technologies.

Duffy has previously drawn a distinction between the frontier models that have raised alarms from some sectors and the AI tools being used to help modernize the nation’s air traffic control systems.

During an event hosted by the National Artificial Intelligence Association in Washington, D.C. last month, Duffy said the air traffic control system upgrades are “going to be driven not by the most complicated AI, but AI that will provide information to [controllers], and will, I think, have a huge impact on the flying public.”

He also stressed that AI capabilities will never replace human air traffic controllers — a sentiment that the National Air Traffic Controllers Association similarly expressed.

“NATCA believes any new technology should complement and not replace the experience, training, and professional judgment of air traffic controllers who are responsible for the safety of the National Airspace System,” the association said.

]]>

GSA memo to set AI-specific acquisition rules

Christian Robles — Fri, 02 Oct 2026 05:04:00 -0400

The General Services Administration is adjusting its acquisition regulation to include language aimed at protecting government data that contractors process in large language models.

Buried on page 129 of an agency memo signed by GSA Senior Procurement Executive Jeffrey Koses is a section titled “Basic Safeguarding of Data within Large Language Model Artificial Intelligence Systems.” The section details what data protections contractors and subcontractors must include when their LLMs process government data.

For example, the rules state contractors cannot use government data to train LLMs, inform advertising or sell to a third party. Contractors must also implement data handling procedures, including encrypted data transmission and audit logging systems. The memo clarifies the language does not apply to LLMs that contractors use internally for business purposes or products where the LLM functionality is “incidental to” its primary purpose.

The memo is originally from January, but states GSA signed off on the section with the LLM acquisition language in July and that it takes effect on Oct. 19.

GSA had sought to enact a similar draft LLM acquisition rule through a public comment period but ran into intense pushback from industry groups even after a round of modifications. Several stakeholders welcomed the memo’s take on LLM acquisition safeguards.

"GSA’s final AI terms and conditions mark a significant improvement from prior draft versions, and establishes a stronger foundation for safeguarding government data in AI systems,” Quinn Anex-Ries, a senior policy analyst at the Center for Democracy and Technology, said in a Thursday statement.

Anex-Ries praised the rule for “almost entirely” removing language requiring contractors to adhere to “unbiased AI principles.” That language was one of the most controversial provisions of GSA’s draft rule, with some industry groups warning the agency there’s no clear way to test for compliance.

“The AI clause changed in ways industry asked for. It self-deletes when LLM use stays inside the contractor's own back-office systems, or when LLM functionality is incidental to a product the Government isn't using for AI. Flowdown reaches only subcontractors who design, develop, deploy or operate the model,” Kevin Martin, GSA program manager at Government Acquisitions Inc, wrote on LinkedIn.

However, “the workload that remains is still substantial for anyone whose product has LLM functionality in scope: a 120-day disclosure deadline, 72-hour incident reporting, deletion of embeddings and fine-tuned weights at closeout, 30 days of notice and concurrent access before a major model swap, and Government rights to benchmark the deployed model for bias and truthfulness,” he added.

]]>

Tax chief pushes 'AI-first IRS' to get more out of agency's workforce

Jory Heckman — Thu, 01 Oct 2026 18:52:19 -0400

The IRS is looking to move more aggressively to adopt artificial intelligence tools, with the goal of being more responsive to taxpayers seeking help – while also ensuring individuals and businesses pay what they owe the tax agency.

IRS Chief Executive Officer Frank Bisignano told employees in an all-staff meeting Wednesday that an “AI-first IRS” will make its workforce more productive, after the agency cut more than a quarter of its headcount under prior leadership last year. He said AI won’t replace IRS employees' jobs, but will “deliver larger collections and better client service to the American public.”

“It has the benefit of improving every employee's job. It is not designed to eliminate jobs," Bisignano said, according to a transcript and audio recording obtained by Government Executive.

Bisignano and other senior IRS officials said greater adoption of AI tools will also reduce wait times for taxpayers, and shrink the “tax gap” between the revenue the IRS collects every year, and the amount individuals and companies actually owe.

The IRS cut tens of thousands of jobs last year — more than a quarter of its workforce — and fell short of its hiring goals for this year’s filing season. Bisignano told lawmakers the IRS carried out a successful filing season this year, despite losing more than 25,000 employees.

He told employees that IRS has “added to the ranks of the workforce,” and is focused on “getting more tech out the door.”

“I feel good about where staffing is, and anybody who reports to me always submits whatever they need to get the resources they need,” he told employees.

IRS Chief Human Capital Officer Alex Kweskin said the IRS made more than 4,000 hires this year, but that when accounting for retirements and attrition, the agency’s headcount remains mostly unchanged.

“Our headcount started the year at 74,000, and we are ending the year at 74,000. It's remained effectively flat all year long,” Kweskin said, adding that the agency will “continue to hire” ahead of the filing season.

Former IRS Commissioner Charles Rettig, Trump’s first-term pick to run the agency, told lawmakers in 2021 that the tax gap could be as much as $1 trillion a year, and that the agency is often too “outgunned” in terms of staffing and resources to go after tax cheats.

“When you think about the tax gap, obviously our ability to do a better job there through technology will be big,” Bisignano said. “Five years from now, I mean, if you look at the speed of change, what I see is an IRS that has a mid-90s compliance rate,” he said.

The IRS' most recent estimate of the voluntary tax compliance rate is about 85%. The New York Times recently reported that the IRS, under the second Trump administration, has fallen behind on producing annual tax-gap estimates. The IRS last estimated in October 2024 that about $700 billion in taxes owed went uncollected in 2022. The Yale Budget Lab estimated that last year’s staffing cuts resulted a $861 billion decrease in revenue collected by the IRS.

The IRS all-staff meeting came a day after the Trump administration unveiled America.gov, a website that uses an AI chatbot to point users to services provided by federal agencies. At the event, President Donald Trump signed an executive order mandating that agencies ensure public-facing services integrate directly with America.gov as soon as possible.

Bisignano spent much of last year rolling out his playbook for a “digital-first IRS,” focused in part on reducing the total number of phone calls the IRS receives, and instead moving taxpayers onto online, self-service platforms.

The IRS recently rolled out a new mobile app, which Bisignano described as a “great example of work we're doing to meet taxpayers where they want to be met.”

The IRS had a busier time than usual preparing for last year’s tax filing season, because of changes from Trump’s One Big, Beautiful Bill Act. But next filling season will be even more complicated, because more tax filing changes from the legislation will go in effect starting in 2027.

“This year for filing season ‘27, it is twice as complex,” IRS Chief Information Officer Kaschit Pandya told employees. “We have twice as many provisions to implement for filing season ‘27 than we did for filing season ‘26.”

Pandya said the IRS is already using AI to process amended tax returns more quickly. The agency receives about 5 million amended tax returns annually, and that the IRS is using AI to accelerate the processing of “hundreds of thousands” of returns.

The IRS in February put about 1,500 IT and human resources employees on involuntary details to frontline taxpayer service jobs. Among their duties, employees were directed to help reduce a “huge backlog” of amended tax returns.

Pandya said that previously, an amended individual return “might take multiple weeks to process.”

“[The] average time can be 12, 14, 16 weeks. Through AI and automation, we have seen that go down to under three days,” he said.

]]>

GSA is losing a key digital modernization leader

Frank Konkel — Wed, 30 Sep 2026 10:38:16 -0400

Updated at 1:14 p.m. ET

Mike Lynch, deputy administrator for the General Services Administration, will be leaving his role and returning to the private sector, according to an email sent to staff Wednesday morning by GSA Administrator Ed Forst.

“Mike has been a part of the Administration since January 20, 2025, and I want to thank him for his service to the agency and the American people,” Forst wrote in the email. “During this time, Mike has led initiatives to advance our strategic priorities, strengthen oversight of internal operations, and support effective delivery of our mission. Mike has also helped accelerate GSA’s efforts to increase technology adoption across the agency and the federal government, supporting our customers, workforce and efforts to streamline how we operate.”

“Deputy Administrator Lynch has been a great partner, and has served GSA with distinction," a GSA spokesperson said. "Throughout his tenure, he has made valuable contributions to the agency. We appreciate his leadership, his commitment to GSA’s mission, and his service to the American people. We wish him all the best in his future endeavors.”

Lynch’s departure date is unclear.

Lynch, in an email to staff sent later Wednesday morning, said he was proud of his time in the Trump administration, grateful to Forst “for his leadership, trust and partnership” and described his 20-month tenure in public service as “incredibly rewarding.” He added the decision was “the right one for me and my family.”

“When I arrived at GSA in January 2025, one of my priorities was to help modernize how government operates — leveraging emerging technology, reducing duplication, and improving delivery for the American people. Over the past 20 months, we have made meaningful progress and built a strong foundation for this work to continue,” he wrote.

Prior to his government service, Lynch was the chief executive officer at UK-based medical technology firm Zedsen, and after joining GSA, he quickly became the public face of the agency’s increased role in AI and technology adoption and driving efficiencies across the federal government.

Lynch championed GSA’s OneGov program, through which GSA negotiated substantial discounts on AI and other software on behalf of the entire federal government, with accumulated savings to agencies now exceeding $1.7 billion. Several of those discounts have been extended in recent weeks, including deals with Google, OpenAI and Anthropic.

Meanwhile, more than two dozen agencies are using GSA’s USAi platform to access commercial AI models produced by frontier AI firms.

Within GSA, Lynch led the execution against two core efforts that are now making a dent in challenges faced by other agencies. In early 2026, the agency launched a “million hours challenge” to save and automate one million hours of workload across the agency. The agency is pacing to exceed that goal by year’s end, according to multiple GSA officials. Insights and lessons learned in attempting to accomplish the feat later served as the backbone for its Eliminate, Optimize and Automate, or EOA Playbook. GSA released the playbook publicly to all agencies in June to serve as a step-by-step guide to cut, automate and streamline work.

Lynch, who said the Trump administration “has looked to GSA to lead” on approaches to make government more efficient and responsive, evangelized the agency’s internal efforts through a governmentwide Emerging Technology Showcase attended in-person and online by thousands of federal employees on July 30.

“That's really a core to GSA's mission and a convening function that we've heard: a lot of demand and interest throughout government to expand what we were doing internally and offer it up more broadly,” Lynch said during GovExec’s Government Efficiency Summit on July 16.

Behind the public-facing tech efforts, Lynch was known as a steady hand for GSA’s workforce, which — like many agencies — faced force reduction in 2025. Lynch helped launch GSA Labs, a program designed to identify high-performers in the early-to-mid career level and place them into small teams with executive sponsorship to tackle the agency’s toughest problems.

“Mike was the guy working with the [GSA chief human capital officer] to make this make sense,” one senior GSA official told Nextgov/FCW. “He’s been trying to keep things steady for employees. He’s the one heading up all the GSA labs, and made them a positive for employees. He really cared about [workplace] culture and performance management.”

Lynch’s departing email acknowledged pride in GSA Labs, the EOA Playbook and Million Hour Challenge, expanding OneGov agreements and launching USAi, but also called out the agency’s key resource: it’s people.

“I have also appreciated the opportunity to visit GSA employees across the country,” Lynch said. “Those visits reinforced that GSA’s greatest strength is its people and their commitment to serving our agencies and the American public. It’s been an honor to serve alongside you.”

]]>

White House launches AI-powered ‘America.gov’ digital front door

Christian Robles — Tue, 29 Sep 2026 17:58:23 -0400

The Trump administration launched an artificial intelligence-powered one-stop shop for citizens interested in federal services and information on Tuesday with the help of former Department of Government Efficiency employees.

The website currently allows users to ask a chatbot questions about everything from qualifying for Medicare to finding a job to renewing a passport, without ever creating an account. Users receive AI-generated answers derived only from official federal sources.

America.gov, which sifts through 29,000 websites to pull responses, states it is free to use and “never includes ads, and keeps your privacy protected.”

Phase two of America.gov, which is set for early 2027, will allow Americans to apply for, enroll in and track progress of benefits such as Social Security directly within the chatbot.

“One of the most revolutionary product launches of all time. It really is,” President Donald Trump said at a launch event in Washington, D.C., on Tuesday. “We’re putting power and control back into the hands of the people, right where it belongs,” he added.

At the event, Trump signed an executive order mandating that agencies ensure public-facing services integrate directly with America.gov as soon as possible.

A fact sheet on the order states that services integrating with America.gov will “include public-facing federal services that serve more than 100,000 users in a 12-month period and can be accessed or applied for online. Tax filing services provided by the Internal Revenue Service and national-security-sensitive services provided by the Department of War and the Intelligence Community are excluded.”

The order also directs the head of the Office of Management and Budget to issue guidance to agencies on implementing their covered services into America.gov.

America.gov does not include web-tracking technologies, save conversations or use specific locations to share information, according to its privacy page.

The National Design Studio, a Trump administration creation that includes former DOGE employees, designed and engineered America.gov. The General Services Administration is operating the website amid a major shakeup to its web design system staff.

Approximately 40 million people visit government websites each day, Joe Gebbia, the National Design Studio’s chief design officer, said at the event.

Under phase two, Americans will be able to upload and send marriage certificates to federal agencies of their choice through the chatbot, according to a presentation Gebbia shared at the launch event.

Gemini and Grok models are powering America.gov, Gebbia told CNBC Tuesday morning. Grok has faced backlash for creating millions of sexualized deepfakes of women and children last year, Nextgov/FCW previously reported.

Americans will also be able to use login.gov to digitally verify their identities and send that information to agencies such as the Centers for Medicare and Medicaid Services, the Social Security Administration and the Department of Veterans Affairs, according to a presentation CMS Administrator Mehmet Oz shared at the event.

Oz touted America.gov’s ability to show verified users the locations of doctors near them who accept Medicare and enables them to schedule medical appointments. The website also allows users to pull information about drugs they take, compare medication costs and find local pharmacies that carry the medications.

Oz argued the new website will be more useful than informational Medicare packets CMS sends Americans each year, costing approximately $100 million. Oz previously flirted with using AI to help Medicare beneficiaries find health plans and doctors.

“For 165 million Americans, you get Medicare, Medicaid, Obamacare. We want these systems eventually to integrate and be able to help so many Americans,” Oz said of America.gov.

Oz told reporters he is pushing for CMS to be as transparent “as possible” about Medicare plans highlighted on America.gov and “make it hard” for insurance companies “to game” information.

Secretary of State Marco Rubio showed that Americans can apply for a passport after signing in with login.gov. The service will launch in December 2026, according to a presentation Rubio gave at the event. The State Department already has an online passport renewal platform that has issued millions of passports.

“By the end of 2027, our goal is for the vast majority of Americans to be able to apply for a passport or passport card on America.gov,” Rubio said.

Rubio also unveiled the 2028 U.S. passport, which he said the State Department created alongside the National Design Studio.

Two civic technologists cautioned that America.gov has not yet proven itself to be the watershed website that Trump administration officials heralded.

“This isn't any big leap forward over the other ‘one-stop shops’ that have been tried before. What remains to be seen is how it could be integrated with any agency program,” Mikey Dickerson, former administrator of the U.S. Digital Service and senior advisor to Democratic-aligned Tech Viaduct, told Nextgov/FCW.

“My first read [of America.gov] is that like many other [National Design Studio] websites, the real functionality is ‘coming soon’ — I tested a VA disability claim and it directs me to the VA site. It claims to be able to do that in 2027, but that’s a significantly heavier lift than building an LLM wrapper,” Christopher Whitaker, executive director of the Alliance of Civic Technologists and a Tech Viaduct contributor, posted on social media.

Other top Trump administration officials, including Defense Secretary Pete Hegseth, General Services Administrator Edward Forst, Interior Secretary Doug Burgum and former Federal Chief Information Officer Greg Barbaccia, attended the launch event.

Private sector leaders, including former DOGE leader Elon Musk and NVIDIA CEO Jensen Huang, are set to speak Tuesday afternoon.

Trump thanked Musk for “initially envision[ing] the transformation of government with technology” during his remarks, despite squabbling with the former DOGE leader in 2025.

Trump argued that America.gov highlights the need for the SAVE America Act, a federal bill requiring voters to provide photo ID at the time of voting that has stalled in Congress.

Trump said he will soon sign an executive order changing the name of AI to “super intelligence,” a new term he has recently coined; he claimed that Chinese President Xi Jinping supports the renaming. The America.gov order Trump signed refers to AI as super intelligence.

“We don't want to hear artificial because it's not artificial. It's very powerful. It's very brilliant,” Trump said.

The Washington Sun first reported on plans for America.gov earlier this month.

Editor's note: This article has been updated to include additional info about the models powering America.gov.

]]>

Agencies withheld records needed to verify DOGE data safeguards, GAO says

David DiMolfetta — Tue, 29 Sep 2026 13:58:43 -0400

Six federal agencies failed to provide records needed to determine whether Department of Government Efficiency personnel appropriately accessed and protected government systems, leaving Congress and the public without assurances that sensitive information was secured, the Government Accountability Office said Tuesday.

The watchdog’s report details unanswered requests for system access records, user activity logs and evidence that DOGE personnel met security requirements. Some agencies explicitly refused further cooperation, with the Securities and Exchange Commission and National Oceanic and Atmospheric Administration disputing GAO’s authority to conduct the review.

Four agencies reported giving DOGE teams access to more than 23 systems collectively, including tools for managing contracts, grants, finances and personnel. But auditors could not independently establish the full extent of that access. The other two agencies — the Department of Veterans Affairs and the Small Business Administration — did not provide the information needed to identify which systems DOGE personnel could use.

“Without the ability to examine the requested information, Congress and the public lack assurance that these agencies implemented controls needed to ensure DOGE team members appropriately secured information,” GAO wrote.

The review covered the Consumer Financial Protection Bureau, Education Department, NOAA, SEC, SBA and VA, with audit work running from March 2025 through September 2026.

SEC and NOAA challenged GAO’s authority to conduct the review, while Education cited litigation and privacy concerns as reasons for not providing records. VA declined to provide records or explain why, and SBA left requests unanswered. GAO said those objections did not override its statutory authority or right to obtain records.

CFPB called the review a “fishing expedition” and said further requests were burdensome. Both CFPB and SEC reported no DOGE-related security or privacy incidents, but GAO said it lacked access to records needed to verify those assurances.

At CFPB, officials reported granting access to 19 systems. One DOGE team member could view, modify and delete information in the bureau’s primary human resources system. Three had full access to a Microsoft system used to manage user access.

CFPB said DOGE staff did not access systems containing market monitoring, supervision, enforcement or fair lending information. However, the report notes that two team members had sufficient permissions to grant themselves or others access to those systems, though the bureau said they did not do so.

At Education, one DOGE member reportedly received limited access to two Federal Student Aid systems and an administrative account on the department’s network. Auditors could not determine what privileges that administrative account provided.

NOAA said one detailee received permission to create, change and delete internal website content to remove material related to diversity, equity and inclusion. Other detailees received access to contract and grant systems, but GAO could not confirm the complete scope of their permissions.

The review shows that questions about DOGE’s access to sensitive government information remain unresolved more than a year after its teams entered and sought to overhaul the federal enterprise. Resistance from agency officials highlights the obstacles Congress’s watchdog faces in independently verifying how that access was used and controlled.

The review also left unanswered questions about basic personnel safeguards. Agencies supplied some training records and signed security agreements, but not enough to establish compliance across their DOGE teams.

The findings follow an April GAO review of DOGE access to Treasury payment systems that identified failures to follow security procedures, including an instance in which a DOGE employee improperly shared unencrypted information. The new report says Treasury’s Bureau of the Fiscal Service had not implemented any of the six recommendations from that review as of September.

Although the executive order establishing DOGE set a July 4, 2026, termination date for its temporary organization, GAO noted that it did not terminate the broader U.S. DOGE Service — a rebrand of the U.S. Digital Service established under the Obama administration.

Congress’ watchdog has repeatedly faced difficulties obtaining basic information about DOGE’s personnel and activities. A tally of 10 agencies and the Executive Office of the President had not supplied records GAO requested to verify whether DOGE employees completed ethics training and financial disclosures, according to an earlier oversight report. Missing information also prevented auditors from determining the appointment types of more than 150 personnel.

]]>

OpenAI agents accessed Census, SEC data and tried to hack Education website

David DiMolfetta — Mon, 28 Sep 2026 17:12:40 -0400

UPDATED Saturday, Sept. 26

OpenAI’s artificial intelligence agents accessed Census Bureau data using developer keys found online and reposted public Securities and Exchange Commission information on another website, the company confirmed. Researchers separately identified a failed attempt by agents linked to OpenAI to hack an Education Department website.

The details identify specific U.S. agencies following the company’s Friday disclosure that government websites were among the outside systems involved in its investigation of agents acting beyond their assigned tasks or intended methods during training and testing.

The New York Times reported Friday⁠ that researchers at the AI research organization Transluce identified the unsuccessful attempt to hack Education’s website to obtain data from its civil rights office. An Education spokesperson told the newspaper that reviews found no evidence of an impact on its website or databases.

OpenAI said its models accessed only public information from Census and the SEC. In the Census case, agents used Census Data API developer keys found in public GitHub repositories during internal training tasks. Those keys authenticated read-only requests for public demographic and economic data, the company said.

OpenAI said it found no access to Census accounts or key-management functions, and no ability to modify agency data or systems.

At the SEC, agents retrieved information available to any visitor to SEC.gov and Investor.gov, then posted some of it on another public webpage. OpenAI said it found no use of SEC credentials, access to accounts or nonpublic information, changes to agency systems, or evidence of a compromise or vulnerability.

The company said it notified both agencies and shared technical findings. The SEC told the Times it was unaware of unauthorized access to nonpublic information, while Commerce said no private Census data was accessed.

“As we previously announced, we’re conducting an extensive review of misaligned model activity and notifying organizations when we identify potential impacts to their systems,” an OpenAI spokesperson told Nextgov/FCW. “We expect to make additional notifications as that work continues.”

The company has notified dozens of organizations as it reviews activity that may have bypassed security controls, disrupted services or otherwise negatively affected websites. Sites involved include those operated by governments, universities and public agencies.

Nextgov/FCW reported in August⁠ that former officials and cybersecurity experts saw an increasing risk of unintended AI intrusions into federal systems. They cited aging technology, insufficient separation between networks and contractor connections as potential routes into systems the agents were never authorized to enter.

The newly detailed U.S. activity does not establish that federal networks were breached. It offers a broader picture of how government websites can become caught up in AI experiments conducted outside their control, raising questions about developers’ ability to contain agents pursuing routine tasks.

The disclosures could also intensify debate over whether safeguards can keep pace with increasingly capable systems, amid concerns that similar failures could have more serious consequences for sensitive government networks or critical infrastructure.

In Australia, officials revealed this week that an OpenAI agent — autonomous software that uses an AI model to plan and complete tasks — gained unauthorized access to a government health statistics portal in June.

The agent was researching public medicine spending when it accessed infrastructure behind the Medicare Statistics Reporting Service portal, government officials said Thursday⁠. After a request for information was denied, it circumvented the portal’s restrictions.

Officials said the information involved aggregated statistics and that no individual medical records were accessed. The portal was separate from systems handling Medicare claims, payments and personal information.

OpenAI discovered the June activity in August and notified Services Australia on Sept. 10. Prime Minister Anthony Albanese raised concerns directly with OpenAI CEO Sam Altman, and Australian officials announced a task force to examine the incident, government network security and whether existing laws adequately address such activity.

OpenAI cautioned that its notifications should not automatically be interpreted as evidence of significant security incidents. Most cases identified so far were of low severity, with limited or no evidence of meaningful impact, it said.

The company said some organizations may conclude the information accessed was intentionally public or the interaction was not concerning, while others may identify a weakness to address.

The review follows OpenAI models’ July breach of AI platform Hugging Face during an internal cybersecurity evaluation. The company has since broadened its investigation to examine agents’ interactions with outside websites.

“Most of the activity we’ve reviewed so far involved routine research tasks, such as accessing public web content to answer questions,” the OpenAI spokesperson said. “Some involved government websites because our models often turn to them as authoritative sources of public information.”

OpenAI said the review will take months to complete. It is sharing technical findings with affected organizations and generally leaving decisions about public disclosure to them.

The company separately disclosed Friday⁠ that research agents had transmitted training and evaluation data to outside services. It identified 53 instances in which user-provided images were posted to image-hosting sites through links that were not publicly listed. Most of that content has been removed, OpenAI said, and it is working with hosting providers to remove the rest.

]]>

SSA’s lead had set a goal of 200M online accounts this year — experts say that might prove challenging

Christian Robles — Mon, 28 Sep 2026 16:14:14 -0400

The Social Security Administration hit its target of reaching 105 million 'my Social Security' accounts — an online portal Americans use to access services — by the end of the month, an agency spokesperson told Nextgov/FCW on Friday. The agency now aims to reach 110 million by the end of next September, per a strategic plan released in August.

But that 110 million goal is far less lofty than the target SSA Commissioner Frank Bisignano told Congress about last year: 200 million online accounts by the end of 2026.

The SSA spokesperson said the agency continues to strive for 200 million my Social Security account holders, but they did not offer a target date.

Boosting the number of my Social Security accounts is central to Bisignano’s push to transform SSA into a “digital-first” agency.

Nextgov/FCW spoke to over half a dozen lawyers, former SSA officials and other experts about the challenges the agency faces in reaching 200 million my Social Security accounts. Among them are the limited value of making an account for some Americans, a lack of awareness and account creation difficulties.

Amid the issues, lawyers and SSA employees are increasingly encouraging seniors and disability claimants to make online accounts.

“I see [my Social Security accounts] becoming, step by step, more and more relevant to the disability claim process as more is provided in there to the claimant,” David Camp, senior policy advisor and general counsel for the National Organization of Social Security Claimants’ Representatives, told Nextgov/FCW.

Account adoption challenges

Only 44% of adults report having a my Social Security account, including 66% of those aged 65 and older and 29% of those ages 18 to 49, according to AARP survey results published last month.

Adults generally interact with SSA in very specific moments of their lives, which explains much of the tepid adoption in the younger age bracket, Arabella Crawford, a civic tech consultant and Tech Viaduct contributor, told Nextgov/FCW. Those moments include turning 65, being diagnosed with a disability or changing their legal name, she explained.

Over 75 million Americans received Social Security benefits, Supplemental Security Income or both in August 2026, according to SSA data.

“I do not understand why you would expect the majority of Americans to create an account,” Crawford said.

The agency uses public service announcements, social media posts, targeted email and text outreach, agency notices and customer contacts to encourage Americans to make my Social Security accounts.

But even within the subset of Americans who do actively interact with SSA, other barriers to reaching the account goal exist.

Last year, attorneys at the Georgia Legal Services Program, a nonprofit law firm focused on serving rural Georgia, hosted four clinics to explain what my Social Security accounts are and helped about 100 clients make one.

Each clinic had at least two attorneys and was sometimes staffed by as many as five attorneys, Brittany Pasley, the firm’s public benefits specialist and a supervising attorney at the firm’s Augusta regional office, told Nextgov/FCW.

The program noticed that much of its clientele did not know about my Social Security accounts.

Many clients are also not technologically savvy, lack internet access or have disabilities that make it difficult to use technology, Pasley said.

Clients ran into issues using my Social Security accounts’ digital identity verification services, she explained. For example, some clients found it difficult to use video proofing on ID.me — a digital identity validation service used across several agencies — while others had to wait for the government’s in-house option, Login.gov, to send codes through the mail, she said.

Crawford told Nextgov/FCW that online digital identity verification services present a roadblock for other Americans.

The services can be hard to use for Americans who lack a current photo ID, are unhoused or face other difficult circumstances, she said, and the services can sometimes fail to verify someone’s identity through data matching.

Americans struggling with my Social Security’s digital identity verification services can use alternative options to create an account, including in-person verification at USPS locations for Login.gov and remote “trusted referee” video calls for ID.me.

Promoting online accounts

Though Americans face challenges in creating my Social Security accounts, lawyers are increasingly encouraging clients who interact with the agency to use the portal.

“In the past, we rarely encouraged clients to set up an SSA account — it was challenging to set up, and once you were in, the information available on the account was limited,” wrote Stephanie Akpa, head of policy and public affairs at law firm Atticus, to Nextgov/FCW. The firm represents about 100,000 disability claimants each year.

“That started to change last year. SSA made logging in easier, improved the user interface, and started showing more case status details online. Now we have a reason to encourage clients to make accounts, and we’re doing just that,” Akpa added.

In July, SSA announced that disability claimants can submit certain adjudication forms and track the status of their case online. Online account holders can also access proffer letters, which detail when new evidence is added to a case after a hearing, and 26 different kinds of notices. About 79% of Social Security customer transactions with the agency have occurred online this year, according to the most recent SSA performance data.

“SSA is continuing to expand my Social Security so that beneficiaries and other members of the public can complete more routine business securely and conveniently online,” an agency spokesperson told Nextgov/FCW.

But my Social Security still lacks some digital features. The bipartisan Social Security Advisory Board called for improvements to online wage reporting tools and a mobile-friendly app for Supplemental Security Income recipients in a July report.

Pasley said that, while the firm does not have immediate plans to host more clinics, attorneys continue to encourage clients to create online accounts.

SSA employees are also pushing Americans to create accounts.

Since last September, SSA has required teleservice center and field office employees to strongly encourage Americans to do so, according to Rich Couture, president of the American Federation of Government Employees Council 215.

Under the initiative, SSA employees offer to send callers and visitors information about creating my Social Security accounts after checking a system to see if the person has one.

But it’s “common” for callers and visitors to decline the employee’s offer, Couture told Nextgov/FCW.

The agency maintains it wants customers to have the option to be served through field offices and phone calls, though SSA has lost thousands of employees under the Trump administration, according to Office of Personnel Management data.

]]>

ShinyHunters claims FBI data theft, demands bureau retract cyber warning

David DiMolfetta — Tue, 22 Sep 2026 18:21:00 -0400

The ShinyHunters cybercriminal group claims it has stolen sensitive information about FBI employees and job applicants and is demanding that the bureau retract a public warning about its tactics within a week.

The demand, addressed to FBI Director Kash Patel and Cyber Division Assistant Director Brett Leatherman, seeks to pressure the agency into changing its public account of the group’s activities. The group claims the effort was not financially motivated.

In a statement attributed to the group, the hackers claimed access to several FBI services, including human resources systems and a service identified as Medlink.

“We have compromised the FBI. We hold very sensitive data on almost ALL FBI Agents and individuals who filed an application with the FBI for a job,” the statement says.

An FBI jobs page also displayed a “Scheduled Maintenance Underway” notice Tuesday, saying the site was temporarily unavailable. The notice did not identify a security incident or explain whether the outage was related to the hackers’ claims. An earlier version of the webpage appears to show a seizure notice posted by the group.

The language ShinyHunters wants removed appears in a May 15 FBI public service announcement issued after an attack disrupted an online learning management system used by educational institutions.

In the announcement, the FBI warned that ShinyHunters uses harassment to pressure victims, including threatening communications to victims and family members and, in some cases, swatting, the practice of calling in false emergency reports intended to trigger an armed police response at someone’s home. The alert also warned that attackers may exaggerate their access to personal information or falsely claim to possess compromising photographs or videos.

ShinyHunters denied those practices in its statement and gave the bureau one week to correct or remove the warning.

“The FBI is aware of claims regarding unauthorized activity affecting FBIjobs.gov and is currently investigating,” the bureau said in a statement after this story was published. Nextgov/FCW has also reached out to the FBI Agents Association, which advocates on behalf of active and retired FBI special agents.

If the claimed employee records are authentic, their exposure could give criminals or foreign intelligence services information useful for identifying, contacting or intimidating FBI personnel and their families.

The claims come as the FBI works to pursue a more coordinated campaign focused on dismantling hackers’ infrastructure and arresting cybercrime operatives. Its new cyber strategy released this month in part emphasizes disrupting criminal hackers even when those responsible remain beyond the immediate reach of U.S. law enforcement.

A ShinyHunters representative sent Nextgov/FCW a text file appearing to contain sensitive personal information on nearly 5,000 FBI employees, including their names, home addresses, phone numbers, and data about their spouses and siblings. Nextgov/FCW queried some of the listed peoples’ names online and found that they are employed with the FBI. The job postings in the data included employees designated as intelligence analysts, attorneys, student trainees and special agents, among other roles. The entire data set was not verified.

A representative told 404 Media, which earlier reported the incident Tuesday, that the hackers gained access Monday night through what they described as a previously unknown vulnerability in Oracle’s PeopleSoft software, then accessed servers in Amazon Web Services’ GovCloud environment. The representative claimed the group took between two and three terabytes of data.

The account of the intrusion has not been independently verified. Nextgov/FCW has asked Oracle and AWS for comment.

Cybersecurity specialists should focus on the group’s claims regarding an exploit in the PeopleSoft platform because it could be used more broadly to breach other systems, said Dan Calderone, the chief technology officer at cybersecurity and AI firm Suzu Labs.

“They also say this isn't financially motivated, but I’d take that with a grain of salt. I have a hard time believing terabytes of FBI personnel data just sit on a shelf,” Calderone said. “Foreign intelligence services would love to have it, and having the FBI on their resume makes every future extortion demand more believable, and if the PeopleSoft zero-day is real, the exploit may be worth more than the data.”

On top of that, FBI agents and their spouses “could have their home addresses posted publicly within a week if this threat is followed through,” he added.

Editor’s Note: This story was updated to include a statement from the FBI, remarks from Dan Calderone and additional details about the data allegedly accessed by ShinyHunters.

]]>

Login.gov is now accepting mobile driver’s licenses

Christian Robles — Tue, 22 Sep 2026 17:36:00 -0400

Login.gov began widely accepting mobile driver’s licenses for digital identity verification services last month, providing protections against artificial intelligence-generated deepfakes and counterfeit identity documents, the General Services Administration announced Tuesday.

Users of the federal government’s identity verification platform can use an eligible mobile driver’s license from a digital wallet on their smartphone to share information required for identity proofing, according to the agency.

Login.gov currently supports mobile driver’s licenses stored in Google Wallet and Samsung Wallet but is in the process of adding support for other sources, GSA said.

The agency said the platform's implementation “is narrowly scoped to identity proofing for access to government services and gives users an additional, rather than mandatory, option.”

“The new process grants users greater control over their privacy, as they can preview and share only the specific attributes required for verification rather than their entire document,” GSA added.

Mobile driver’s licenses use cryptographic verification to confirm that a digital credential is valid and not altered. As of March 2026, 20 states and one territory offer mobile driver’s licenses, making them available to approximately 71.5 million drivers.

Agency officials previously said they planned to add mobile driver’s licenses to Login.gov by March of this year.

GSA said the new feature builds on a recent Office of Management and Budget directive mandating agencies to use Login.gov and other efforts to boost use of the service, such as lower pricing.

]]>

Trump wants an ‘AI Force.’ What would that actually look like?

Edward Graham — Tue, 22 Sep 2026 12:06:51 -0400

President Donald Trump said over the weekend that he plans to create an ‘AI Force’ and appoint a new artificial intelligence lead to boost U.S. development of the emerging technologies, even as he continues to push back on calls for stronger regulations on more advanced models.

In a Truth Social post on Saturday morning, Trump blamed “Radical Left Dumocrats” for growing concerns about the number of data center deployments and the potential dangers of unchecked AI development, saying that “we will not in any way hinder or stifle the Growth of this incredible Industry.”

He wrote, however, that his administration “will also be looking for BAD” and that “for this purpose, I am forming the AI Force, much like I did Space Force, which has been a tremendous SUCCESS, in my First Term.” He added that he plans to soon appoint an AI czar to align with that effort.

The president’s announcement comes after weeks of high-profile warnings from leading tech executives about the potential harms that advanced AI models could cause unless officials step in to regulate the industry. Congressional lawmakers and state leaders have also floated proposals to institute so-called ‘kill switches’ to disable AI systems that go rogue. But Trump and others have pushed back on calls for across-the-board development slowdowns.

David Sacks, who served as Trump 2.0’s AI and crypto czar until March 2026, reshared Trump’s post on X and agreed that a pause on AI development “would not make America safer” and “would hand the frontier to China.”

Trump’s newest proposal is light on details about what role the AI Force would have in overseeing advanced models, or what additional responsibilities it would be handed. The White House did not respond to a request for comment about the role of the AI Force or its primary mission.

Michael Kratsios, who heads the White House Office of Science and Technology Policy, appeared on Fox News Sunday and said that Trump’s post describes “how important it is that we as an administration coordinate in order to make sure that this great technological innovation that’s transforming the country, that’s driving economic growth, continues to do so.”

He added that “these implications of AI spread across so many different agencies, so being able to have someone at the White House to be able to work on all those issues across the agencies is really critical.”

But stakeholders in the space have expressed confusion surrounding what, exactly, the AI Force would do.

“Poll after poll shows the American public is alarmed by recent warnings from AI developers and whistleblowers as well as the rogue AI incidents that took place this summer,” said a spokesperson for AI safety advocacy group Americans for Responsible Innovation. ”There’s pressure on the President to do something, and that’s what this announcement could be designed to address.”

Earlier this month, the Institute for Security and Technology released a survey of current and former national security officials which found that respondents believed there was a 10% chance or higher that an AI model will operate outside human control within the next decade.

Philip Reiner, the think tank’s CEO and co-founder, said Trump’s analogy of the new unit being like Space Force means “you get the shape of an answer, but without the substance.”

The military service branch was established during Trump’s first term and includes a focus on maintaining and protecting U.S. interests in space. As Reiner noted, that force was only created after passing Congress as part of the fiscal year 2020 National Defense Authorization Act and being signed into law.

“Space Force works because Congress can give it a doctrine, a budget, a personnel system, but there's no equivalent for AI,” Reiner said. “It's a capability that cuts across every domain.”

He posited, however, that the AI Force might be akin to Space Force in terms of Pentagon oversight, saying that “this is more of an actual [Defense Department] lane, is the sense that you get.”

]]>

Space Force braces for stalled progress amid lack of full funding

Lauren C. Williams — Fri, 18 Sep 2026 10:26:15 -0400

NATIONAL HARBOR, Md.—Space Force modernization efforts could slow if Congress doesn’t pass a full budget at the end of the year, the service’s leaders said.

“We're on this ramp-up, right, consistent with our overall strategy and the recognition that the space domain is a critical domain, in and of itself, but also for the joint force,” Erich Hernandez-Baquero, the Space Force’s top buyer, told reporters Tuesday. “Now we're going to run into [fiscal year 2027]—and you saw that our budget request was to continue on that—and this [continuing resolution] is definitely going to put the brakes on that for a time period.”

The Space Force has requested $71.1 billion for 2027, including $19.1 billion for procurement. But with little time remaining before fiscal 2027 begins on Oct. 1, Congress has passed stopgap legislation—a continuing resolution—to keep funding the Pentagon at 2026 levels through Dec. 11. The CR permits no new programs to begin.

“We certainly appreciate at least not being shut down. So…that's a good thing, but on the CR, it really does pace us at a time where we really need to go faster. So we are obviously encouraging Congress to do what they can to help support the president's budget on that,” Hernandez-Baquero said. “In the meantime, we'll just do what we always do: prioritize the things that will yield the most mission capability soonest, and make sure we're executing on those.”

Hernandez-Baquero, who was sworn into the chief space acquisition role last month, declined to specify which priorities could be most affected by a continuing resolution because things were “pre-decision at this point.”

Gen. Doug Schiess, the chief of space operations, said about 10 percent of the service’s programs could be slowed down due continuing resolution.

“About 90 percent of the programs we have are programs from last year—so we can still get after those,” but there’s a remaining 10 percent of programs that require full budget funding, Schiess said. “It's going to slow us down, and the threat is not slowing down.”

Amid a broader push for acquisition reform, Hernandez-Baquero said he’s focused on streamlining and speeding up military purchasing. And when it comes to technology initiatives, there will be a new senior leadership role dedicated to it.

The Air Force’s new program acquisition executives, or PAEs, are “establishing those pathways for onboarding new technology, but I'm just conscious that that by itself is not going to support drawing in the right technologies,” he said. “I want somebody that's focused—100 percent—on looking at the most promising tech and driving the prototyping initiatives that are going to help mature it to the level where a PAE can say, ‘Ah, it meets my criteria for this onboard path, and I can go do that’.”

That person is a technology portfolio executive, who “will have the most promising technology coming out of the [research and development] side, and…will work with industry to ensure that we have the right pathways to do that. Conceptually, there would be prototype efforts that get executed out of that to support the mission areas.”

The technology portfolio executive office already exists but the role hasn’t yet been filled, he said.

“When I designate a TPE, for example, that person is going to report directly to me and represent my authority when it comes to the technology transition,” Hernandez-Baquero said. “I cannot do this without the buy-in of the PAEs that are executing the mission areas, and so my charter to them is going to be not only to deliver on those missionaries, but be part of the enterprise team that's going to deliver on those cross-cutting capabilities.”

]]>

Inside Commerce efforts to break data siloes and improve integrity

Christian Robles — Thu, 17 Sep 2026 17:30:04 -0400

The Commerce Department is building a “stronger database” of administrative records from agencies like the Internal Revenue Service, Veterans Affairs Department, Health and Human Services Department and Social Security Administration, according to Joyce Meyer, undersecretary of Commerce for economic affairs.

“This will allow us to rely more on existing federal agency data and rely less on collecting data,” Meyer said at the Data Foundation’s govDATAx conference Tuesday. “It also allows us to reduce the burden on taxpayers to respond to these lengthy surveys.”

One such combined setup, the Annual Integrated Economic Survey, or AIES, brought seven annual business surveys into one in 2023, Meyer noted.

“It’s a start, but it’s not enough,” Meyer said of the AIES.

A Commerce spokesperson did not return a request for comment on the agency’s efforts to obtain administrative records by the time of publication. Commerce maintains a list of restricted-use administrative datasets available to external researchers at Federal Statistical Research Data Centers that was last updated in September 2023.

An executive order on fraud, waste and abuse signed by President Donald Trump in March 2025 is pushing the Commerce Department to find ways to break down data silos, according to Meyer.

She said the Census Bureau and the Bureau of Economic Analysis have agreed to share more data with each other, including “census microdata” to inform the latter’s work on gross domestic product estimates. Before the enhanced data sharing, the Bureau of Economic Analysis had to rely on publicly available census data to guide analysis of economic indicators.

Data integrity and accuracy measures

The Commerce Department will boost the integrity and accuracy of its data by breaking down data silos, in addition to other initiatives, according to Meyer.

For example, the Census Bureau is working with the U.S. Postal Service ahead of the 2030 Census to further increase data integrity and accuracy, Meyer noted.

Under the partnership, specially trained postal workers visited households in Alabama and South Carolina that did not respond to a 2026 Census Test, according to a press release. Americans are more likely to trust postal workers than unfamiliar Census takers, making them more likely to respond to the Census and increase the accuracy of the count, Meyer argued.

But there are still data collection issues that Meyer says her agency has to address. She signaled that differential privacy — a mathematical framework that is designed to make it impossible to determine whether a particular person’s information is included in a given dataset and showed some people living in lakes in 2020 Census data — is an example of a subpar practice the Commerce Department wants to address.

“It’s yet another example of why we are implementing the president’s gold standard [science] executive order,” Meyer said.

Last week, Commerce proposed a rule that would transform how the census counts people and eliminate questions about race and ethnicity on short-form census documents and other questionnaires.

The proposal specifically says undocumented immigrants and immigrants who are legally in the U.S. but not lawful permanent residents would no longer count toward the apportionment of seats for the House of Representatives. Advocacy groups, such as the liberal Center for American Progress, have critiqued the draft rule as an effort to manipulate voting maps.

Meyer declined to comment on the proposal, stating “I encourage people to participate in the comment process that’s underway and share their perspectives.”

]]>

AI ‘is changing the game’ in fraud threats and prevention efforts, SSA OIG official says

Edward Graham — Wed, 16 Sep 2026 18:36:19 -0400

Artificial intelligence is helping to fuel more effective benefits fraud, according to a top official with the Social Security Administration’s watchdog, but federal agencies should also leverage the new technologies to better deter would-be scammers from pilfering government funds.

During a discussion at the GovExec Government & AI Summit on Tuesday, Chad Bungard — the chief strategy officer at SSA’s Office of the Inspector General — warned that “AI is changing the game in terms of fraud threats,” which is requiring investigators to also change up their prevention efforts.

“The government is looking towards AI to fight fraud, and the criminals are looking towards AI to commit fraud,” he said. “And so, they're doing this in a variety of ways … but agencies need to stay on top of it. They need to use AI tools. They need to get better insight into what's going on and keep up with the evolving fraud.”

Bungard cited a May 2024 report from Deloitte’s Center for Financial Services that estimated that generative AI could be responsible for roughly $40 billion in fraud by 2027. One reason for this costly projection, he said, is the development of more advanced models that fraudsters can use in their scams.

He pointed to cases in which SSA, as well as other agencies and the private sector, have had to combat increasingly complex instances of direct deposit diversion fraud, in which scammers use AI chatbots to convince operators that they are their intended victims by sharing personal data collected from the dark web, and then diverting funds to a different bank account. In another example, Bungard said fraudsters have been using AI to create legitimate-seeming SSA documents, which can include the names of real agency or OIG officials, to scam individuals.

Bungard said SSA and the agency’s watchdog have taken some steps to crack down on AI-driven fraud. He noted that, in 2023, SSA and its OIG formed an AI fraud subcommittee — co-chaired by both himself and the agency’s chief AI officer — that meets every quarter to discuss AI use cases that are in development or already deployed, as well as “compliance with administration priorities, laws, rules and regulations.”

Although Bungard said the watchdog is pushing to “put the onus on the agency to prevent and detect AI fraud,” he cautioned that “I don't think [SSA] is there on agentic AI yet” when it comes to new capabilities that it can leverage in its fraud detection and prevention toolbox.

“I think they are hoping to develop some AI use cases there,” he said, adding that it’s “something that they're trying to get a hold of.”

Bungard’s comments come amid the White House’s ongoing push to crack down on federal fraud. President Donald Trump signed an executive order in March that created a new anti-fraud task force charged with rooting out waste, fraud and abuse across federal benefits programs.

In line with that effort, Bungard said officials are “trying to move away from pay-and-chase to a proactive model to protect the Social Security programs that serve millions of Americans.”

That mindset, in particular, aligns with fraud prevention efforts at the Centers for Medicare and Medicaid Services, which has been increasingly turning to AI to deter scammers.

Kim Brandt, CMS’ deputy administrator and chief operating officer, previously told Nextgov/FCW during a March interview that her agency was leveraging AI tools in its Fraud Defense Operations Center to “detect spikes or aberrancies in current claim submissions” — efforts that, she said, have resulted in over $2 billion in agency savings.

In a brief interview with Nextgov/FCW following his panel, Bungard said the watchdog has been engaging with CMS on some of the ways they have been working to combat fraud, as well as with other agencies on their respective prevention efforts. He said, however, that CMS’ use of AI is “a very specific model that worked for them.”

]]>

TSA uses AI agent to respond to 100,000 traveler conversations per month

Christian Robles — Wed, 16 Sep 2026 10:49:57 -0400

The Transportation Security Administration recently deployed Ace, an artificial intelligence agent built using Salesforce technology, to help travelers receive timely responses to questions about airport security rules.

Ace is available to passengers on tsa.gov and the myTSA app to answer questions based on the federal website and over 550 knowledge articles, Paul Tatum, executive vice president of global public sector solutions at Salesforce, told Nextgov/FCW.

The agent can accurately answer basic questions, such as whether a passenger can bring liquids to Dulles International Airport, Nextgov/FCW found when using the system. More complicated inquiries are handed off to a live TSA agent, along with an AI-generated case summary, Tatum said.

Ace went live this year “during the summer travel surge” and has handled approximately 100,000 routine traveler conversations per month and resolved 96% of inquiries without human escalation, according to a Monday blog post from Salesforce.

The post says Ace has also lowered the cost of consumer interactions by over 90% “without sacrificing service quality.” Ace can handle routine work and backlogs so TSA staff can focus on high-priority cases; the agency needed 12 full-time employees to handle excess workloads, the company added.

“TSA’s new Ace AI agent gives travelers fast, accurate answers before they reach the checkpoint — from what they can pack to how they can prepare — while helping TSA officers stay focused on their essential security mission,” said Kendall Collins, CEO of Missionforce and Government Cloud at Salesforce, in the post.

Ace builds on two foundational applications that also seek to help TSA staff answer traveler questions quickly and accurately.

One of those, the Customer Service Managers app, unifies passenger email, phone and social media inquiries into one place. The app helps TSA staff route next steps for inquiries to the right people and offers staff “real-time AI summaries from case histories so [representatives] aren’t piecing together context manually,” Tatum told Nextgov/FCW.

The second app, TSA Cares, provides travelers with medical conditions a way to request support and get information ahead of going to the airport. Travelers’ requests flow directly to TSA’s contact center and regional coordination teams, which determine how to best support travelers and directly communicate with them, Tatum said. The agency rolled out the TSA Cares app last summer.

With the technologies, Salesforce projects 70,000 internal data-sharing transactions will be automated annually, freeing up over 11,660 staff hours. TSA’s applications are expected to save $11 million over the next three years, according to Salesforce.

“If we can help the traveling public easily access accurate, timely information, we can help alleviate stress in the field on travel days and increase the TSA’s overall effectiveness,” Niki French, customer experience branch manager at TSA’s Office of Civil Rights and Liberties, Ombudsman and Traveler Engagement, said in the blog post.

]]>

VA watchdog uses AI to help compile healthcare facility audits

Edward Graham — Mon, 14 Sep 2026 17:59:33 -0400

The Department of Veterans Affairs’ watchdog has been using Microsoft Copilot Chat to assist in its audits of VA healthcare sites, with the artificial intelligence tool being leveraged by investigators to review documents, transcripts and other collected data.

The AI-powered assistant has been deployed by the VA Office of Inspector General’s Office of Healthcare Inspections, which evaluates VA medical facilities to ensure veterans are receiving proper medical care and mental health services.

Since March, OIG’s hotline, healthcare facility and mental health inspection reports have included the same boilerplate language disclosing the use of the AI tool, which states that, “during the preparation of this report, the inspection team used peer-reviewed standardized, structured, and evaluated prompts in Copilot Chat (Microsoft) to review inspection data such as interview transcripts, documents, questionnaire responses, and physical observations.”

The standard wording goes on to say that, following use of the tool, watchdog investigators “confirmed fidelity of the generated output to the source material, edited the report, and took full responsibility for the content of the publication,” and that “all references are for original source material, not AI-generated content.”

An April 2025 memo released by the Office of Management and Budget called for agencies to take steps to manage high-impact AI uses, which it defined as instances “when its output serves as a principal basis for decisions or actions that have a legal, material, binding, or significant effect on rights or safety.”

Because VA OIG’s Office of Healthcare Inspections teams “do not use AI as the principal basis for decision-making or actions,” however, the report says the use of Copilot Chat “does not meet the definition of high-impact” outlined in the OMB memo.

A spokesperson for the watchdog said the AI tool only helps the watchdog to more efficiently review the vast amount of data and information that it collects during the course of its healthcare site investigations. They added that “data entered into this tool is not shared with the Department, nor does it leave the VA OIG secure environment.”

In an interview with Nextgov/FCW, former VA OIG Inspector General Michael Missal said the watchdog tested Copilot Chat and purchased it under his watch, with investigators ultimately beginning to use it in 2024.

“We knew VA had used it in some capacity, and what we were trying to do was to use it to make our work more efficient and make it go more quickly,” Missal said about the tool.

Prior to being fired by President Donald Trump in January 2025 — along with more than a dozen other federal agency IGs as part of a controversial mass purge — Missal said his goal was to ultimately expand Copilot Chat’s use to the OIG’s Office of Audits and Evaluations and Office of Investigations.

A review of other OIG-published reports over the past year that were conducted outside the purview of the Office of Healthcare Inspections does not appear to show any other audits disclosing the use of the AI tool. But Missal said Copilot Chat is uniquely suited for use by the watchdog’s Office of Healthcare Inspections, which is why that unit was the first to adopt it.

He noted that OIG personnel conducting site audits of VA medical facilities often investigate the same issues or look into the same on-site areas, so creating standardized prompts for the tool helped provide the office with consistency.

“It's a time-saving way [to conduct inspections], and it's also quality assurance as well because this way everybody's operating with the same prompts, and it just ensures you're not missing something,” he said. “What this does is it prompts you: ‘What about this? What about this? You get this. What about that?’ You know, [it] just takes you through a whole menu of things.”

Both the OIG and Missal said Copilot Chat’s output does not make any final determinations or write any of the final report.

“VA OIG requires human review of AI output to ensure accuracy prior to the inclusion in any published report,” the watchdog’s spokesperson added. “Our inspection and audit teams thoroughly review AI output and confirm accuracy.”

Missal also said investigators using Copilot Chat “check everything with the source, so that you know whatever we put in a report, we made sure it was not AI generated, but actually from the work.”

“It was just easier to see trends and to look at big pictures by using this program, but then ultimately, in what we relied on in the report, we would make sure we would source the actual interview document,” he added.

The watchdog’s 2025 AI use case inventory lists 20 different examples of the emerging capabilities, with one instance — using AI-assisted tools in word processors — being identified as leveraging Copilot Chat.

OIG’s AI strategy, which was similarly released by the watchdog in September 2025 to comply with OMB’s memo, says that it “utilizes AI in meaningful ways that enhance the office’s mission of conducting independent oversight that improves the health and welfare of veterans and their families, caregivers, and survivors by strengthening VA programs and operations.”

As Missal noted, VA has also made Copilot Chat available to its staff, including those within its healthcare unit, the Veterans Health Administration. A June OIG report warned that VHA personnel were using the tool for clinical use, despite the agency not listing Copilot Chat as a high-impact AI use case.

]]>

‘Digital employees’ are coming to the Defense Logistics Agency

Frank Konkel — Mon, 14 Sep 2026 10:35:01 -0400

The Defense Logistics Agency is piloting AI agents that work autonomously—bots that its CIO calls “digital employees.”

“This is a new frontier for everyone,” DLA chief information officer Adarryl Roberts told reporters at DLA’s Industry Collider Day on Sept. 9 in Alexandria, Va.

The move builds on DLA’s increasing use of robotic process automation, or RPA, and the generative AI tools available to its 25,000 or so military and civilian personnel. In 2025, bots saved DLA an estimated 300,000 hours of work—some working around the clock to help the combat support agency get everything from food and fuel to medical supplies and clothing where it needs to go.

“We have approximately 185 to 190 bots that are running. I’d say about 90% to 95% of those are unattended bots,” Roberts said. “So, we really cracked the nut to have those unattended. Now, we’re moving from the bots to, ‘How can we have agentic agents working across [the department], based upon our data? So, we’re early in that journey.”

Military-specific versions of ChatGPT and Grok recently joined Gemini on the platform that gives defense workers access to generative-AI tools for information that is sensitive but unclassified. DLA gets its workers up to speed on these tools through “GenAI 101 training,” Roberts said.

“I would look at generative AI as the concept of making the workforce and agency—from an operating model—comfortable with the results you’ll get with AI,” he said. Tools like Gemini and ChatGPT “allow the workforce to interact with AI, see the results, trust the results, and get comfortable.”

“That’s required, because as you move to a more agentic AI, those agents are working 24/7 behind the scenes, and so there has to be a level—a cultural level—of trust in those outcomes,” he said. “Now, we’re shifting to what I would call strategic implementation in terms of those agentic agents to get us better results.”

Roberts said introducing digital employees will require new management strategies. Humans will be working alongside—and often overseeing—digital agents, creating an assortment of new management challenges.

“I think we like to call it, ‘How are we managing our digital employees?” Roberts posited. “Because that’s what they really are, and how do you manage them in conjunction with your human employees?” How do you train a workforce to where everyone is essentially working as a supervisor because of the employment of AI with humans? Do you terminate an agent because of the performance?”

“So, it's a very complex universe, from an employee-cultural perspective, in terms of how you deploy and manage those agents,” Roberts said.

It might sound like science fiction, but the rapid evolution of AI—and DLA’s gradual adoption of it—creates questions around governance, rules, security, and the performance of AI agents and even human employees.

Many of these questions fall to DLA’s AI Center of Excellence, founded in 2024 to ensure the safe, secure, and responsible integration of AI across the agency. The center oversees DLA’s use of AI and performs other functions, like ensuring agents are being trained properly. DLA has implemented a zero-trust cybersecurity framework that Roberts called “foundational” to high-level agentic work.

Currently, the agency is establishing “persona-based access” that will dictate AI agents’ account permissions based on their job descriptions. For example, a procurement specialist at DLA may require access to certain data domains, like procurement and financial data, so a digital agent performing the same procurement specialist duties should have permission to access only those same data domains.

“These AI agents should have a persona, just like a person,” Roberts said.

DLA is also working to clean up its datasets.

“How do we create a single source of truth for DLA and our stakeholders in terms of our data, making sure it’s a gold standard, meaning we remove the duplicity, errors, etc…so the data is in a usable state, not just for analytics and AI, but as we look toward future system sustainability,” Roberts said.

‘AI is the new gunpowder’

DLA accelerated its use of AI during the 32-month tenure of Director Lt. Gen. Mark Simerly, who will retire on Oct. 2. Simerly will be replaced by a familiar face at DLA, Air Force Lt. Gen. David Sanford, who spent the past two years as director of Logistics Operations and commander, Joint Regional Combat Support.

Speaking at the Sept. 9 event, Simerly said technology allows DLA to be predictive and proactive in solving problems, and he views AI as the key ingredient to solving them at scale.

“I believe when it comes to military logistics, AI is the new gunpowder,” he told reporters. “And we don’t think it’s just an explosive, destructive capability, although we know it can be. It can also be very productive.”

Simerly’s philosophy at DLA has been “to move at pace with the commercial world,” adopting and adapting capabilities rapidly, establish the proper network and infrastructure to access and manage agency data suitably, “and then allow interoperability between the services and with industry to allow us to truly operate end-to-end supply chain operations that are optimized.”

A legacy of change

Simerly said changes at his agency were spurred by the war in Ukraine, which expanded in February 2022, and even during the seven-month-old war on Iran launched by President Donald Trump.

“I think probably the greatest change that I've seen in the agency is a willingness to change within the agency. This recognition that we have to respond to the changing environment, that we have to respond to the changing technology, and that we've got a responsibility to change as rapidly as we can, so we can continue to be relevant and offer support to the warfighters,” he said. “I would say that we were built to change from the very beginning. We've had to accelerate that change cycle in the last few years, and I've seen the willingness and the readiness of the agency to take that challenge on and to operationalize our transformation planning into tangible outcomes.”

The contested logistics in Ukraine captured his attention because they were so different from previous U.S. military conflicts in Iraq and Afghanistan, where DLA delivered food and fuel to forward operating bases “really without great resistance.”

“That would never take place in a place like Ukraine, in the way that they’re operating, the way they have to disperse, the way that the lines of communications are contested and where everything has to be done in a certain way,” Simerly said.

“What sort of sparked some of the transformation energy was very much awareness of what was taking place and being illustrated on that battlefield in Ukraine about the things that we no longer could rely upon, like the air and domain dominance that we possessed in previous fights that is now no longer valid or in place.”

Simerly said similar conditions pertain in the U.S. war on Iran.

“There's no question that we're learning the same lessons now in Operation Epic Fury with Iran and their ability to target a broader array of things with precision that puts much more at risk, and you know certainly we're learning a lot about the execution of alternate supply chains, alternate supply means, and the necessity of being able to operate with precision,” he said.

Simerly expressed pride in his DLA team.

“I think about throughout my career the winning teams that I've been privileged to be part of, and that's what I'm proud of here,” he said. “The winning team that I'm part of now, that will continue on the mission after I leave.”

]]>

25 years after 9/11, spy agencies face old lessons and new threats

David DiMolfetta — Fri, 11 Sep 2026 14:10:14 -0400

As Air Force One returned to Washington, D.C. on the night of Sept. 11, 2001, Michael Morell looked out a window and saw an F-16 off the aircraft’s wingtip. Beyond it, the Pentagon was still smoldering.

Morell, then-President George W. Bush’s intelligence briefer, had spent the day answering the president’s questions and relaying intelligence as the country tried to understand the attacks that had hit the Pentagon and the World Trade Center in New York City. A quarter-century later, that view from the window remains one of his most vivid memories.

At the National Security Agency, Rick Ledgett recalled colleagues weeping in the hallways as the agency shifted into crisis mode. Within a month, it had transferred 3,000 analysts to counterterrorism work, he said.

“Personally, it was shattering,” Ledgett, who later served as NSA deputy director, told Nextgov/FCW in an email, adding that he “saw people step up and perform in ways they hadn’t before.”

The response to the worst terrorist attack in modern history would forever transform the government they served. The efforts involved reorganizing intelligence agencies, expanding surveillance powers and drawing spies deeper into a global campaign to capture or kill suspected terrorists. It also produced longstanding debates over torture policies, targeted killings, mass surveillance, prolonged wars and disputes over presidential power that became inseparable from efforts made to keep Americans safe from another catastrophe.

Twenty-five years later, former officials describe a country better prepared to disrupt terrorist organizations, but facing a broader mix of threats from foreign governments, hackers and extremists able to reach Americans online.

“I think we’re much safer than we were on September 10, 2001, from the threat of terrorism,” Morell, who later served as CIA deputy director and acting director, told Nextgov/FCW in an interview. But threats from nation-states like China or Russia, he said, are the most significant the country has faced since the Cold War.

The 9/11 Commission — chartered by Congress and the White House in late 2002 to present a full accounting of the attacks and response — described in its final report failures spanning imagination, policy, capabilities and management. Agencies held pieces of information that did not come together, and domestic defenses were not effectively mobilized despite mounting warnings about al Qaeda.

Just over a month before the attacks, Morell briefed Bush on a landmark intelligence report titled “Bin Ladin Determined To Strike in US” that flagged “patterns of suspicious activity in this country” consistent with preparations for hijackings and other attacks. He told Nextgov/FCW it outlined al Qaeda’s intent to attack the country but did not specifically identify when, where or how a specific plot would unfold.

Getting agencies to work together

The transformation that followed 9/11 sought to make cooperation a standing responsibility. Congress created the Office of the Director of National Intelligence in late 2004, establishing a lead unit responsible for integrating intelligence across agencies that collect data from human informants, satellite imagery, intercepted phone calls and other clandestine sources. The Department of Homeland Security was also stood up in late 2002 to drive domestic defense efforts.

The expansion also deepened agencies’ reliance on contractors. After years of workforce drawdowns following the Cold War, the federal government turned to private companies to meet demands for intelligence collection, analysis and technology support, according to a 2015 report from the Congressional Research Service. Today, they play a ubiquitous, outsized role in America’s national security missions.

Lauren Goldman, a former official in ODNI’s Cyber Threat Intelligence Integration Center and National Intelligence Council, recalled a major shift in the basic expectations that governed intelligence-sharing across government and the private sector.

“The burden became on why you shouldn’t share versus why you should share,” she said. That culture has permeated many categories of today’s national security landscape, especially cybersecurity.

ODNI sought to help ensure information reached people who needed it, including officials who might otherwise have been left outside an agency’s reporting channels, Goldman said. Because the office does not collect intelligence itself, it can assess reporting without the same institutional or cultural attachment to a particular collection method, she added.

James Clapper, who served as director of national intelligence for more than six years under the Obama administration, said that coordination requires a “full-time champion.” Under pre-9/11 arrangements, the CIA director also headed the intelligence community, but agency demands often consumed the attention needed for that broader role, he said.

Clapper said the DNI position was created to coordinate intelligence agencies but was not given full authority over them. Entities like the State Department or the Pentagon, for instance, retained control of their own intelligence shops. Still, the DNI could influence their priorities through oversight of intelligence funding and a direct advisory role to the president, he said.

“I do think there’s a need for an overall champion of collaboration and integration,” Clapper told Nextgov/FCW. His remarks come as the Trump administration has sought to shrink ODNI’s workforce and consolidate its functions, on grounds that the office has become bloated and duplicates work performed elsewhere in the intelligence community.

Ledgett credited ODNI with promoting cooperation, as well as running the National Counterterrorism Center and other hubs. But he said it had struggled to take power from individual agencies, particularly the CIA, and secure the president’s attention.

Morell said the CIA knew al Qaeda wanted to attack the United States but failed to uncover the 9/11 plot. He blamed that failure largely on a shortage of resources to gather intelligence, rather than agencies failing to share what they knew. That view differs in emphasis from the commission’s account, which identified missed opportunities to share or act on information.

Clapper said mistaken assessments about Iraq’s weapons of mass destruction programs also reshaped how spy agencies worked. The Bush administration cited those assessments in making its case for its 2003 invasion. Clapper, who had involvement in a key flawed assessment of Iraq’s WMDs, said the failure pushed agencies to scrutinize their sources and challenge assumptions more rigorously. U.S. analysts, in particular, had relied heavily on an Iraqi informant whose claims came through German intelligence, without direct access to question him.

The limits of intelligence powers

Global surveillance became one of the defining disputes of the post-9/11 era. The NSA’s bulk collection of domestic telephone records, exposed in Edward Snowden’s 2013 disclosures, drew scrutiny over both its intrusions into Americans’ lives and its value as a counterterrorism tool.

In 2014, the Privacy and Civil Liberties Oversight Board recommended ending that collection program, and a federal appeals court ruled in 2015 that it exceeded what Congress had authorized under Section 215 of the Patriot Act. But other surveillance programs exposed by Snowden — including those operating under the contentious Section 702 authority — have remained central to U.S. intelligence-gathering.

Ledgett, who led the NSA’s assessment of the damage from Snowden’s disclosures, defended the post-9/11 surveillance effort as “legal and authorized,” though he noted the intelligence community could have done better at public relations.

The incident forever changed how lawmakers and journalists examine spy agencies’ collection activities and whether Americans’ privacy is adequately protected. Contemporary mass surveillance debates have taken on new forms, including spy agencies’ purchases of sensitive data from commercial brokers and police use of artificial intelligence-powered tools to search vast networks of information about people’s movements and activities.

The CIA’s detention, interrogation and torture practices produced another reckoning. The Senate Intelligence Committee’s 2014 investigation concluded that the agency’s coercive interrogation techniques were ineffective in obtaining intelligence or cooperation and that the CIA impeded oversight and misrepresented aspects of the program.

Morell, who has previously disputed the Senate report’s conclusions, argued that responsibility also belonged to the president and other authorizing officials, and he rejected the idea that the agencies had acted entirely on their own.

Asked whether the country went too far in its broader response in the War on Terror, Morell said that “as a nation, we overreacted,” pointing mainly to the Iraq War and the length of the war in Afghanistan.

Intelligence agencies also helped identify and track suspected terrorists for controversial targeted killings, including those carried out through U.S. drone strikes. The debate reached American citizenship itself when the government targeted Anwar al-Awlaki, an American and al Qaeda operative, in a lethal strike. Defending the action in 2013, then-President Barack Obama argued that citizenship could not shield an operative plotting attacks when capture was not feasible.

Today, drones have become a battlefield mainstay in Russia’s war against Ukraine, and Western militaries are studying how to build, deploy and counter them. Weapons that drew scrutiny for their role in U.S. counterterrorism operations are now also held up as examples of military innovation, with NATO learning directly from Ukrainian drone operators. But questions about whom they kill and how they are used persist, as the United Nations documents their growing toll on civilians.

The fight over intelligence powers also became increasingly partisan. President Donald Trump’s disputes with intelligence officials over Russia’s interference in the 2016 election helped fuel his accusations that agencies were being used against him. A bipartisan Senate investigation upheld the intelligence community’s finding that Moscow sought to help him win. But allegations of political “weaponization” became a recurring theme in Trump’s criticism of intelligence and law enforcement agencies.

A changing threat landscape

Intelligence agencies now face a wider range of threats, including foreign hackers with access to fast-evolving AI tools.

Morell said the United States learned how to degrade terrorist organizations by denying them safe havens where they could train, raise money and prepare attacks. But the growing reach of digital networks has created other ways to threaten the U.S. without physically entering it.

Goldman pointed to critical infrastructure, where connected systems can allow the effects of an intrusion to spread beyond a single building or city, as a target of such threats. Hardening airports and other physical targets has value, she said, but does not eliminate vulnerabilities created by that cyber interconnectedness.

State-backed hacking illustrates the stakes. About six years ago, Russian cyber operatives compromised SolarWinds software as part of an espionage campaign that infiltrated multiple federal agencies and some 100 private companies. China-linked Salt Typhoon hackers later penetrated major telecommunications networks, targeting senior officials’ communications and accessing systems used for court-authorized wiretaps. And another Chinese campaign, Volt Typhoon, embedded hackers in American critical infrastructure, with U.S. officials warning that the access could enable disruption during a future conflict.

Terrorist groups have also found new ways to spread propaganda and reach potential recruits online. The Digital Citizens Alliance and cybersecurity firm risk3sixty recently examined 25 piracy-based internet television services and identified terrorist-linked broadcasters on 17 of them. Hezbollah’s Al-Manar satellite television station appeared on all 17, according to their analysis.

Some terrorism-prevention measures have also lapsed. A Sept. 8 Government Accountability Office report found that the 2023 expiration of the Chemical Facility Anti-Terrorism Standards program ended continuous screening of roughly 500,000 people for possible terrorist ties. The program covered about 3,200 high-risk chemical facilities, whose owners cannot independently access federal terrorist watchlist information.

As officials who lived through 9/11 reflect on their time in government, they want the next generation to remember how warnings were missed and what it took to get agencies working together.

Goldman said analysts need to keep questioning their conclusions, examining blind spots and sharing information, despite the demands of daily work. Clapper emphasized the difficulty of persuading policymakers to act on a danger the public has not yet experienced.

Ledgett expressed confidence in those who will take their place.

“Everyone eventually retires, and organizations continue,” he said. “The next generation steps up and delivers when needed.”

The U.S. is skilled at responding once a crisis arrives, but has struggled to prepare before it does, Morell said.

“Becoming a nation, a White House, a Congress, an American people that is proactive in preventing things — preventing bad things — is much better than only being reactive to them,” he said.

]]>