Building AI into the fabric of federal operations

When I sat down last summer with Pete Tseronis for the debut episode of the CGI Voices podcast, I described the attention being paid to artificial intelligence as one of the most amazing hype cycles I had ever witnessed—and not all of the hype was exaggerated.

In the nearly one year since that conversation, it has become increasingly true. The large language models and machine learning algorithms that power generative AI are not new, but they have reached a critical mass, and we are past the tipping point. 

The government is taking AI seriously, both its potential benefits and its risks. In late March, the Office of Management and Budget (OMB) issued Memorandum M-24-10, “Advancing Governance, Innovation, and Risk Management for Agency Use of Artificial Intelligence,” which requires agencies to appoint chief AI officers and details the requirements of the role. 

The memo follows President Biden’s October 2023 executive order on “safe, secure and trustworthy artificial intelligence ,” which expressed the goal of ensuring “that America leads the way in seizing the promise and managing the risks of” AI.

It might seem unusual to see the federal government, historically risk-averse and slow to adopt new technologies, joining in the excitement over AI. I think it is because, in this era of emphasizing customer experience, agency leaders and policymakers recognize the potential for AI-powered tools to improve the delivery of services. That said, security and risk management are critical components in any technology adoption. Smooth, secure, painless and value-enhancing adoption is the end goal of establishing the CAIO role.

Balancing risk and reward

Each agency’s CAIO will become the linchpin of that agency’s AI implementation plans and projects. The CAIO’s challenge will lie in balancing the rewards that well-implemented AI could bring with the security and performance risks that come along with it. Agency leaders should consider these principles as they design their AI programs:  

• Enact secure guardrails. Establish a governance framework early. OMB’s guidance recommends measures to protect the rights and safety of the public, which include conducting AI impact assessments, testing AI in a real-world context, and identifying factors that might contribute to algorithmic discrimination. 
• Establish use cases. Like any technology, AI is the best choice for some applications and not appropriate for others. However, it is transformative enough that some serious business process re-engineering may open pathways to use cases that don’t currently exist. Be creative in rethinking how you do what you do, and whether generative AI might be able to fill a gap.
• Engage industry and government collaborators early. Collaborate with experts before issuing requests for proposals. Involving industry stakeholders early in the strategic planning process allows government entities to benefit from their expertise and insights into emerging technologies. This early engagement facilitates a deeper understanding of the challenges and opportunities in implementing AI solutions, allowing government agencies to make informed decisions, and develop effective strategies for incorporating AI technologies.

That last point, I believe, is particularly important in exploring any emerging technology. Innovation is rarely confined to a single source. By fostering collaboration between government and industry—and in the same vein, between multiple government organizations—agencies can enhance their AI capabilities, drive innovation and address complex challenges more effectively in the evolving technological landscape.

Balancing security with mission is always a delicate matter, and the potential of AI—great harm as well as great improvement—makes it even more difficult. Security functions within the agency must share the commitment to the goal of mission execution, while the mission functions must respect the need for security. 

Learn how CGI can help guide your federal AI journey.