OMB Issues First Governmentwide AI Acquisition Policy

The Office of Management and Budget (OMB) released new guidance today to improve how Federal agencies acquire artificial intelligence (AI) technologies. The guidance, outlined in the memo Advancing the Responsible Acquisition of AI in Government, directs agencies to boost cross-functional collaboration, manage AI risks and performance, and foster a competitive AI market. (MeriTalk October 3, 2024)

OMB’s Deputy Director for Management, Jason Miller, emphasized the need for responsible AI procurement, stating that Federal agencies will either have AI systems built by contractors or purchase them directly. “This new memo equips agencies with the tools to capture AI’s potential while managing its risks,” Miller said. (ibid)

A large portion of the memo focuses on managing AI risks, with OMB mandating early involvement from agency privacy officials in AI acquisition processes to identify privacy risks and ensure legal compliance. Agencies are also instructed to negotiate contracts that require vendors to provide detailed information for evaluating AI systems, assessing risks, and protecting government data. (ibid)

The guidance addresses generative AI specifically, calling for testing, red-teaming, and evaluation to ensure the safety and appropriateness of AI tools. It also promotes practices to avoid vendor lock-in, prioritize transparency, and ensure interoperability in AI systems. (ibid)

This guidance fulfills a key part of the Biden-Harris administration’s October 2023 AI executive order and reflects input gathered from public comments and industry roundtables. OMB’s Miller highlighted the Federal government’s significant purchasing power, noting that in 2023, it spent over $100 billion on IT products and services. He stressed that responsible procurement decisions can accelerate AI advancements while mitigating risks for government use. (ibid)

Questions concerning the new OMB issued AI guidance? Give us a call.