4 Ways Agencies Can Successfully Embed AI Into Operations

When the terms “AI” and “government” are uttered in the same sentence, they tend to trigger a mixed response. Some people are optimistic, realizing the potential for AI to create efficiency and opportunity across the public sector. Others are more resistant, acknowledging that biased algorithms have the potential to hurt — not help — government agencies and the citizens they serve. 

Both of these responses are valid. AI is full of complexities, and government leaders are still wrapping their heads around the challenges and opportunities that come with it. But, when implemented effectively, AI has the power to make government operations more efficient, freeing up time and energy for staff to focus on more strategic initiatives and better serve constituents. 

On our first episode of the AI Tipping Point Podcast, Government Executive Media Group President Constance Sayers sat down with guests representing the public and private sector to discuss the double-edged sword that is AI. The general consensus? Successful AI adoption in government is possible, but it will require strict governance, more resources and consistent communication. 

Here are the steps that can help government agencies get there:

1. Understand Your Mission

When agencies undergo any sort of digital transformation, it’s usually to solve a specific challenge. Before agency leaders can effectively implement AI, they must determine the problem they’re trying to solve. 

“Start with the mission,” said Jamie Milne, Senior Engagement Manager at World Wide Technology. “Think about what's going to be most impactful for your bottom line. AI is really just a means to an end to perform your mission better.”

Once agency leaders identify these challenges and understand how they fit into their organization’s larger mission, they can determine the technology that is best suited to meet their needs. 

One public sector organization leveraging AI to solve agency-specific challenges is the U.S. Secret Service. Ryan Moore, assistant special agent in charge in the U.S. Secret Service, and one of the guests on the podcast, said the agency is looking at opportunities to use machine learning and natural language processing to solve its enterprise challenges. 

“We have an immense amount of data, and so we’re starting to look for opportunities to generate greater insights so we can achieve cost efficiencies within our own operations,” he said. 

Over the next year, Moore hopes to integrate NLP into the agency’s service desk operations. He also wants to use AI to generate greater insights from financial data to drive more strategic decision-making. 

2. Change the Way You Look at AI

For many agency leaders, the thought of implementing an agencywide AI adoption strategy can seem like a herculean effort requiring a complete overhaul of technology and processes. However, Milne noted that oftentimes agencies can take a much simpler approach. 

“There are many different ways to solve these challenges,” he said. “There’s a kind of aura and mysticism around AI. Many customers think about AI and jump from elementary math to deep neural networks and image processing and video analytics. But there seems to be very little appreciation for what goes on in between, which can often be much easier to understand, much easier to implement and much more accepted by the workforce.” 

Milne said instead of setting large — and often unrealistic — goals for adopting AI, agencies are better off taking small steps to integrate AI into their operations gradually. Ultimately, by changing the conversation and simplifying the applications of AI, agencies can embed these tools more efficiently. 

3. Acknowledge the Ethical Challenges — and Opportunities — of AI 

Once an organization has identified an AI adoption strategy, it can begin to work to communicate that strategy internally. 

Agency staff may have several questions and concerns when it comes to AI adoption. For example: Will jobs be replaced by machines? Will AI be able to track and analyze personal information? How exactly will AI be leveraged across the organization? 

According to Jonah Hill, the U.S. Secret Service’s senior cyber policy advisor and former policy specialist at the Department of Commerce’s National Telecommunications and Information Administration, the ability to answer these questions — and create internal policies to regulate the use of data — is key to public sector AI innovation. He added that policymakers have a critical role to play in how agencies use AI, but warned against implementing too many regulations. 

“There is a rush, often, to regulate technology in a way that can really hamper innovation,” he said. “So it’s important for policymakers and government officials to step back and say, alright, what are the things that we need to really think about when it comes to AI?”

Answers to many of these ethical questions can be found in guidelines and considerations released by the Organization for Economic Cooperation and Development in 2019, which looks to serve as a document government officials can turn to as they formulate AI policy. For example: AI should benefit people and the planet by driving inclusive growth, sustainable development and wellbeing; it should be designed to respect the rule of law; and organizations and individuals developing or deploying AI should be held accountable for the proper functioning of those systems. 

As more agencies adopt AI on a larger scale, public policy will need to evolve with them. In the meantime, agency leadership should communicate clear policies internally to help staff make smarter AI decisions. 

“Governance is really the unsung hero of all of these AI projects,” Milne said. “Projects with the greatest success are built upon a clear foundation of data governance and management.” 

4. Identify Gaps in the Workforce — and Close Them

The Secret Service has taken measures to implement a data governance structure agencywide by identifying leaders who can enforce these policies. A chief data officer leads strategic planning initiatives, while a chief data architect applies domain-level knowledge of enterprise tools and data structures. Moore said this internal expertise — both strategic and technical — has been key to establishing effective data governance.

However, agencies must also be proactive about identifying talent at every level across the enterprise. According to Milne, building a multi-discipline team is often challenging because many organizations operate in silos. Instead, cross-team collaboration can help spark innovation. Agencies should identify future leaders and implement learning and development programs to help them gain the skills necessary to keep up with the pace of innovation. With the right people and policies in place, organizations will be well-equipped to tackle AI innovation — now and in the future. 

“We need to do a better job managing cross-functional talent and find connective tissue between silos,” he said. “We need folks who can manage and configure the deep technical components. Then they need to communicate with the data scientists and the folks who are implementing the algorithms to leverage that technology. There also must be open communication between those folks and the people who are actually deciding how to apply the data science across the organization.” 

To learn more about how government leaders can embrace the AI revolution, check out the full AI Tipping Point podcast.