President Trump signed a much-anticipated Executive Order (EO) establishing a framework for the federal government to review AI frontier models for cybersecurity risks. The EO calls on AI companies to voluntarily share models for testing by federal agencies 30 days before public release. The review will focus on cybersecurity and national security, but the order does not establish specific licensing, permitting, or safety rules. The full text of the EO can be found here. Coverage can be found on ABC News, CyberScoop, and Federal News Network.
Thank you for reading! Below, we’ve rounded up additional tech policy news nd shared a featured podcast.
This Week in Washington
- FedScoop: The National Institute of Standards and Technology (NIST) relaunched its AI Safety Consortium as the NIST Artificial Intelligence Consortium. The new group, which includes 280 partners, will focus on the science, evaluation, and testing of artificial intelligence tools and models.
- The Hill: The U.S. Department of Commerce closed a loophole in a 2023 AI chip export rule, clarifying that a license is needed to sell advanced AI chips to both Chinese companies and companies with Chinese parent companies. The department has previously said it plans to replace the rule entirely, but has not yet released a plan.
- The Record: The National Security Agency (NSA) selected acting chief David Imbordino as the full-time chief of the Cybersecurity Directorate. Imbordino is a veteran NSA staffer who is currently the acting director of the department and previously served as NSA’s Election Security Lead for the 2020 election.
- Nextgov: The General Services Administration (GSA) plans to reform its AI acquisition rules to increase the government’s preference for fixed-price purchases of AI software. The full rule will reform AI contracting practices across the federal government and is expected to be released this month.
Article Summary
- CNN: Microsoft and the Mayo Clinic announced a partnership to integrate medical data, research, and physician expertise into an AI model specifically for clinicians. The multi-year project will also aim to build a patient-facing healthcare AI assistant.
- Tampa Bay Times: With hurricane season officially beginning, the National Hurricane Center is planning to use AI models alongside traditional monitoring for the second year in a row. The Center’s annual review found that AI models generally outperformed other models in predicting the strength and path of hurricanes last year.
- NBC News: Illinois passed landmark AI safety legislation, joining California and New York in enacting a framework for AI transparency and accountability. Among other provisions, the bill requires AI companies to create plans to address catastrophic risks from their models and submit to annual independent audits of their models on safety issues.
- Associated Press: Generative AI is creating new access to music creation. The Associated Press reported on one musician, Samuel Smith, who has been able to create new music with artificial intelligence even after his Parkinson’s disease took away his ability to play guitar.
Featured Highlight
- Tools and Weapons with Brad Smith
Microsoft’s quarterly AI Diffusion report has been updated for spring 2026. Microsoft President Brad Smith hosted AI for Good Lab Director Juan Lavista Ferres to talk through the findings and what they mean. They focus specifically on the urban-rural divide in American AI adoption and how AI is changing software development and knowledge work. (Special Edition: Tracking AI diffusion in Q1 2026 – May 28, 2026 – 31 minute listen)
Note: Voices for Innovation regularly shares a range of opinion articles and press releases from organizations in and publications covering tech policy. These pieces are meant to educate our audience, not to endorse specific platforms or bills.