AI: Year in Review and Expectations for 2024

It is no surprise that artificial intelligence (AI) significantly increased in scale and implementation in 2023. With the introduction of generative AI (ChatGPT) by the start of 2023, alongside actions taken by agencies to establish sound data governance and infrastructure, AI was bound to shift the federal environment on several fronts this past year.

In January 2023, NIST released the final version of the AI Risk Management Framework (RMF). The RMF outlined the characteristics of a trustworthy AI system as well as key functions to mitigate risks in AI systems already in practice. The RMF set the stage for federal change in culture and perspective on the design of AI systems and monitoring of AI risk.

Throughout the year, Congress proposed a series of legislation to establish an AI governance framework and accountability measures to advance AI innovation while minimizing negative impacts of the technology. Proposed legislation ranged from establishing commissions to oversee separate aspects of AI development, transparency, and fairness, to defining standards for AI-generated content identification and upskilling the federal workforce. Congress held numerous briefings, as well as AI insight forums, to form an understanding of and aid in AI legislation development.

The White House, for its part, held multiple meetings with leaders in the industry to examine specific aspects of AI. The White House also secured a set of commitments from several leading AI organizations on the safety, security, and trust of AI systems.

Federal agencies, such as the State Department, released inaugural AI strategies in 2023 to guide organizational AI adoption and safety. CISA also released an AI roadmap to both promote the use of AI to enhance cyber capabilities and ensure AI systems are protected by advanced threats. At DOD, the department updated its Data, Analytics and Artificial Intelligence Adoption Strategy to accelerate the adoption of AI capabilities to warfighters and decision makers.

2023 federal AI activities culminated in late October with the release of the highly-anticipated AI executive order from the White House. The comprehensive Executive Order (EO) 14110, Safe, Secure and Trustworthy Development and Use of Artificial Intelligence, provides guidance and actionable steps to facilitate and expand federal use of AI through eight principles such as new standards for AI safety and security, advancing equity, and promoting innovation and competition. The EO delegates multiple new tasks and initiatives to agencies and prompts legislative action for responsible AI innovation.

Shortly after the release of the AI EO, OMB issued a draft memo to federal agencies on the governance, risk management and innovation of AI technologies. Specifically, the memo details prioritization of IT environments to enable AI application and streamlining future AI investments while providing sound oversight of the technology.

Prompted by many of the provisions on the AI EO alone, AI is expected to be a major federal tech trend in 2024 and beyond. Here are some thoughts on top AI trends contractors can expect in 2024 (among numerous other AI trends, I’m sure):

• Continued release of AI policies and guidance. The release of the NIST RMF, AI EO and OMB memo mark the beginning of legislation and policies expected for AI, and in particular generative AI.  In May 2023, OSTP issued an RFI forecasting the development of a National AI Strategy. NIST recently released an RFI seeking comments to develop guidelines in areas such as AI red-teaming and generative AI risk management.
• Hiring and upskilling of federal AI staff. The AI EO and OMB memo call for the installation of Chief AI Officers and AI governance boards at top federal departments. However, agencies will look to continue to try to grow and train an AI workforce base, particularly through federal tech talent programs, to safely adopt and maximize use of the technology.
• Expansion of agency generative AI adoption. Federal agencies are grappling with how to safely integrate generative AI into operations. While some are openly exploring generative AI, others have put blocks on the technology, such as SSA. Nonetheless, the AI EO discourages general bans and barriers of generative AI use, and agencies are likely to begin exploring generative AI use, as seen by recent actions at Education.
• Increased use of AI to bolster cyber capabilities. Not only will federal scrutiny increase over the security of AI offerings, but agencies will need to incorporate AI in cyber defense activities, particularly with the growth of AI-powered threats. 
• Use of AI to “sift through the noise.” Data is the building block to leveraging AI, however, automation and large language models in activities such as data classification and tagging will see increased investment to help agencies sort through volumes of information.
• Continued modernization of IT infrastructure to support AI. Agencies will focus on making existing systems and networks AI-ready. The OMB memo calls on agencies to develop data infrastructure to process datasets used for testing and operating AI solutions. IT leaders will also focus investments in areas of compute, storage, and system interoperability to deliver AI solutions.
• Standardization of AI acquisition methodologies. GSA is proposing a government-wide buying capability, such as a blanket purchase agreement or acquisition vehicle, to vet and standardize federal purchase of AI.