The field of artificial intelligence is evolving rapidly, and the U.S. government is working to adapt to these constant changes as quickly as possible.
To ensure that AI is brought in successfully, federal organizations have devised new strategies to guide their adoption of AI tools. During a panel discussion at the Potomac Officers Club’s 5th Annual AI Summit last week, experts shared their thoughts on how new approaches to AI implementation are defining the use of these technologies in the public sector.
Dr. Alex Jaimes, chief AI officer at Dataminr, said that many major transformations in core AI technologies occurred several years ago, and what is happening now is “that now there are many, many more larger models that are available.”
What these developments have done, he continued, is significantly break down barriers to AI and allow users to “do a lot more with less knowledge.”
Echoing his thoughts, Nick Weir, head of federal delivery at Yurts, said that what drives AI evolution today is the ability to combine technologies to conceive more powerful capabilities.
Integrating different tools with large language models, he said, “is really transforming how people use these models.”
Dr. Amanda Bullock, AI lead at the Air Force Research Laboratory, noted that this increased accessibility has enabled those without a background in AI to take advantage of these technologies.
Federal organizations have recognized the increasing availability of AI, and many agencies are now offering programs to strengthen their AI workforce to leverage these cutting-edge capabilities.
The Department of Homeland Security has made commercial generative AI available to “any employee that takes training and gets approval,” DHS Chief Technology Officer for AI and Emerging Technology Chris Kraft shared during the discussion.
Bullock said the AFRL is promoting AI exploration for its workforce using three key methods: hands-on training, internship programs and knowledge sharing.
One key trend characterizing today’s AI landscape is the concept of responsible AI (RAI), which focuses on the ethical considerations surrounding AI implementation. The October 2023 Executive Order on the Safe, Secure, and Trustworthy Development and Use of AI set the direction for the U.S. government’s RAI direction and now guides agencies as they integrate AI into their operations.
Kraft said the Executive Order “put a good framework in place” to fill in gaps that exist in AI regulations. He expects more related policies to materialize in the near future.
Federal agencies and contractors alike have already embraced the standards set by the Executive Order. One way they are doing so is through workforce education, which has shown to be a major factor in the responsible use of AI.
To keep up with new AI technologies as they emerge, Jaimes said Dataminr hosts internal sessions to make sure everyone is on the same page about these developments. He said that when ChatGPT was released, the company began these sessions within a week.
According to Kraft, DHS has launched training programs on bias and other potential shortfalls of generative AI and has also established responsible use working groups.
Another component of RAI is trusting the user to follow the proper guidance on AI use, Bullock said.
Kraft pointed out that while AI has become a prominent topic within the U.S. government, it is also a global phenomenon.
“It’s not just federal, the Executive Order and what the states and locals are doing, but it is the EU AI act and how that worked through the system and what impact that is going to have,” he said.
The Potomac Officers Club’s renowned Annual CIO Summit will return for the 5th time on April 17. At this engagement, top GovCon CIOs and other information technology experts will gather to discuss the current challenges and opportunities shaping the federal IT landscape. To learn more and secure your spot before tickets sell out, click here.
