4 Ways Government Agencies Can Prioritize AI Adoption This Year

In 1958, Former President Dwight D. Eisenhower determined that the American government was falling behind on scientific innovation after the then-Soviet Union famously launched the first human-made object into space. His solution was to create an agency that would set the United States up to win the tech race. The agency, which became known as the Defense Advanced Research Projects Agency (DARPA), funded research projects and fostered public-private partnerships that led to technological advancements. Just years after its creation, DARPA began experimenting with programs that would pave the way for AI innovation across the public sector. 

Fast forward six decades and the federal government has witnessed a number of new innovations thanks to increased data storage and processing capabilities. Agencies are realizing the endless possibilities of AI — and increasing their investments accordingly. In 2018, DARPA announced it would allocate $2 billion toward programs that aim to push the boundaries of AI across the Department of Defense. The White House is also on board: In 2019, President Trump issued an executive order announcing the “American AI Initiative,” a strategy guiding government leaders in the investment and advancement of AI.

Moreover, AI is predicted to save the government billions of dollars annually. And according to a report from Deloitte, investment in AI could free up government workers’ time by 30%. 

So where do we go from here? With so much potential and so many possible use-cases, many government agencies don’t know where to start when it comes to AI adoption. On our recent episode of the AI Tipping Point podcast, Government Executive CEO Tim Hartman sat down with Mike Madsen, director of strategic engagement for the DoD’s Defense Innovation Unit (DIU), and Anthony Robbins, VP of Federal at NVIDIA, who have seen these challenges firsthand. Guests agreed: 2019 was the tipping point for AI, but 2020 will be the year that it makes its way into operations. 

1. Bridge the Gap Between Concept and Application 

For too long, public sector leaders have been talking about what an AI-driven government might look like, but making strides toward adoption is easier said than done. 

“There are hundreds of AI pilots happening across the Department of Defense,” Robbins said. “But to really make an impact, we have to move from the labs and the prototypes to the actual operations. That’s when we can really start adding value to our missions and better support citizen services.” 

Turning a concept into reality is no easy feat, but agencies can begin the journey by identifying use-cases where AI can play a role.

“As we bring new projects and programs online, we need to start by considering how we’re going to bring in AI,” Madsen said. “Identifying these opportunities is the first step toward harnessing the power of artificial intelligence.” 

The DoD announced in January it will begin using AI for predictive maintenance. DIU has partnered with C3.ai to implement software that would help predict repairs on military aircraft. The tool would improve the availability of aircraft for military missions. 

DIU is also looking at how AI and machine learning algorithms could help deliver supplies to survivors of natural disasters. If successful, this use of AI could also assist in the fight against COVID-19 by providing masks and other materials to hospitals or sending tests to communities. 

2. Create an Organizational Culture That Embraces Change

Before agencies can carry out successful AI adoption, they need to foster an organizational culture that supports innovation. That change starts from the top. 

To promote agency-wide innovation, Madsen said leaders should not only embrace failure, but encourage it. He has witnessed the DoD bring these ideas to life with Kessel Run, the Air Force’s in-house software lab. 

“We took a concept of DevOps where we were writing code right alongside the end user and created a software that spawned off into Kessel Run,” he said. “We weren't necessarily prototyping a technology, we were prototyping a methodology, and it resulted in a new way of doing business.”

3. Reevaluate Your Talent Management Strategy

Successful AI adoption is driven by people — and it’s only effective with the right personnel on board. Identifying key internal stakeholders can empower leaders to drive these initiatives forward. This is what Robbins calls the “magic of middle managers.” 

“Many of these internal stakeholders have been at your organization for five or 10 or 20 years,” he said. “They own the budget, they own the systems in place that have to change. So it’s important to work with the middle managers on a transformation agenda that can set agencies up for success.”

In the long term, it will also become important to generate new data-centric career paths that will ultimately help emerging leaders develop skills that prepare them to navigate an AI-driven future. 

4. Treat AI Adoption as a Team Sport 

There’s no doubt that the road to AI adoption is a long one, but agencies don’t need to walk it alone. By leveraging additional expertise from academic institutions and industry, agencies can access a number of opportunities to accelerate AI adoption efforts. 

For example, agencies can enable career mobility between the public and private sectors by introducing rotational programs in which private sector talent can spend some time working in government organizations or projects. The result is two-fold: Agencies gain expertise and insight from world-class talent, while staff gain experience working on different types of projects. 

According to Robbins, industry leaders also have a role to play in educating the public sector market. His company, NVIDIA, hosts events and continuing education programs to teach government employees skills and facilitate the exchange of best practices. 

As AI continues to accelerate at record speed, these partnerships will enable government agencies to lead the charge. 

“Everybody plays a role in this journey,” Robbins said. “It’s a team sport. And if we play our position right, we can all help the government move forward faster.”

To learn more about what it takes to make AI a success in government, check out the AI Tipping Point podcast and NVIDIA.