UPDATE — SEPTEMBER 2025: The U.S. AI Safety Institute (U.S. AISI), housed at NIST, has expanded from an initial ~280 members to over 400 consortium participants, bringing in a wider mix of civil society, academia, and smaller labs alongside Big Tech. In mid-2025, AISI began publishing draft test suites for large language model (LLM) safety, covering robustness, content risk, and alignment benchmarks, now open for public feedback. Agreements with leading AI companies such as OpenAI, Anthropic, and Google DeepMind have formalized structured federal access for pre-deployment model testing.

On the regulatory front, the Commerce Department’s proposed rule requiring quarterly safety test reporting under the Defense Production Act advanced into the public comment stage, with officials signaling a possible final rule in early 2026. Meanwhile, NIST’s content labeling challenge yielded new watermarking and provenance techniques, and the White House is weighing their integration into federal procurement this fall.

Internationally, the U.S. hosted the first International Network of AI Safety Institutes conference in San Francisco (Nov. 2024). By mid-2025, the network included counterparts from the EU, UK, Canada, Japan, and Australia, with biannual coordination meetings now in place. The U.S. will host the second global conference in Washington, D.C. in spring 2026.

The National Security Memorandum on AI (Oct. 2024) is now being operationalized: the Department of Defense, Intelligence Community, and DHS are piloting joint AI red-team exercises with AISI. In parallel, the U.S. Patent and Trademark Office (USPTO) launched a second consultation on AI disclosure and copyrightability in invention filings, with final guidance expected in 2026.

Finally, the $100 million AI-enabled semiconductor research competition announced in 2024 has moved forward—awardees were named in July 2025, with projects beginning this fall, underscoring Commerce’s commitment to secure compute and innovation capacity.

ORIGINAL NEWS POST:

U.S. Commerce Department and NIST Mark One Year of Progress with New Safety Initiatives and Global Partnerships

The U.S. Department of Commerce is celebrating significant progress toward several AI goals. Secretary of Commerce Gina Raimondo marked the anniversary by highlighting the Department’s achievements in implementing the order’s directives, just ahead of the first International Network of AI Safety Institutes conference.

Since a presidential executive order was signed in October 2023, the Commerce Department has launched the U.S. AI Safety Institute (U.S. AISI), a national body dedicated to AI testing and safety. The institute is tasked with driving collaboration across sectors, including partnerships with AI developers and national security agencies. To date, U.S. AISI has formed a consortium of approximately 280 members and conducted pre-deployment testing for new AI models.

The release of a National Security Memorandum (NSM) on AI last week further underscored the administration’s commitment to integrating AI into national security. This memorandum designates the U.S. AI Safety Institute, based within the National Institute of Standards and Technology (NIST), as the core agency for AI model testing, ensuring collaboration with national security and intelligence agencies to advance responsible AI adoption.

The executive order has spurred the Commerce Department to implement a series of critical actions, including the following highlights:

• AI Safety Testing and Evaluation: The U.S. AISI initiated safety testing agreements with leading AI companies, supporting pre-deployment evaluations of major AI models.

• Regulatory Oversight: The Commerce Department proposed a rule mandating AI developers to report safety test results and other critical information quarterly, using the Defense Production Act to enforce these standards.

• Transparency and Content Labeling: The U.S. Department of Commerce submitted a comprehensive report to the White House detailing methods for labeling AI-generated content and improving transparency, including NIST’s challenge to develop new techniques for detecting AI-driven content.

• Guidance on Dual-Use Foundation Models: Commerce released an extensive report on the benefits and risks associated with dual-use AI models, including policy recommendations. This report was shaped by feedback from hundreds of stakeholders.

• International Collaboration: Secretary Raimondo and Vice President Harris unveiled U.S.-led initiatives to promote AI safety at international events, including the 2023 UK AI Safety Summit. The administration plans to launch a global network of AI Safety Institutes, set to meet for the first time in San Francisco this November, to harmonize safety standards and policies globally.

• Intellectual Property Guidance: The U.S. Patent and Trademark Office published guidelines on AI-assisted inventions and AI technology’s role in patent applications, ensuring that the evolving field aligns with intellectual property laws.

• Support for Sustainable Semiconductor Research: A $100 million competition was launched to drive AI-enabled research for developing sustainable semiconductor materials and processes.

Need Help?

If you have questions or concerns about the U.S. Department of Commerce, or any global, AI reports, guidelines, regulations and laws, don’t hesitate to reach out to BABL AI. Their Audit Experts can offer valuable insight, and ensure you’re informed and compliant.