The Office of Management and Budget (OMB) has released Memorandum M-24-18, introducing new requirements and guidance for federal agencies on the responsible acquisition of artificial intelligence (AI). This memorandum aims to harness AI's potential to modernize agency operations and enhance government services, provided that the risks associated with AI technology are mitigated. Importantly, this memorandum will influence industry standards as AI-related products and services conform to this guidance.

Building upon previous efforts like the Advancing American AI Act and Executive Order 14110, the memorandum directs agencies to improve their capacity for responsible AI acquisition by focusing on three key areas:

Cross-Functional and Interagency Collaboration: Agencies are required to "create or update acquisition policies, procedures, and practices to reflect new responsibilities and governance for AI," as established by OMB Memorandum M-24-10. This includes sharing information on AI acquisitions across the executive branch to promote consistency and cooperation.

Managing AI Risks and Performance: Recognizing that "AI poses novel types of risk," the memorandum supplements existing risk management requirements. Agencies are instructed to adjust their acquisition practices to account for "the distinctive ways that AI systems and services are developed, trained, and deployed." This involves implementing practices to ensure effective risk management for rights-impacting and safety-impacting AI, addressing issues related to privacy, security, data ownership, and interoperability.

Promoting a Competitive AI Market with Innovative Acquisition: To ensure agencies can "always procure state-of-the-art AI," the memorandum requires prioritizing decision-making that enhances interoperability and prevents vendor lock-in. It also encourages the use of innovative acquisition practices to support a diverse, competitive, and resilient federal AI marketplace.

By proactively addressing risks and fostering collaboration and innovation, the OMB aims to ensure that AI technologies are acquired and implemented in ways that benefit government functions and public services while safeguarding against potential harms. You can access the new guidance using the link below.