Trump AI Order Sets Voluntary 30-Day Federal Review Of Powerful Models
On Tuesday, June 2, 2026, President Trump signed an executive order asking AI companies to voluntarily share powerful new models with the federal government before public release.[1]
The order asks firms to submit their most powerful models for testing up to 30 days before public release, down from a 90-day review in an earlier draft.[1] It directs federal agencies to create benchmarks for assessing models' cyber capabilities and to set up an AI cybersecurity clearinghouse to review and share information about vulnerabilities.[1] The order also says the voluntary testing framework does not authorize mandatory government licensing, preclearance, or permits, and says binding regulation would have to come from Congress.[1]
The White House delayed a planned signing last month over concerns that a 90-day review could stifle U.S. AI innovation and weaken American firms' position against China.[1] Administration worries intensified after Anthropic said in April it limited release of its Mythos Preview model because the system could identify and exploit software security vulnerabilities.[1] That announcement helped spur the push for federal benchmarks and the new cyber clearinghouse.[1]
CBS characterized the program as giving the government an "early look" at powerful models under a voluntary system.[2] The administration's compromise on timing reflects an attempt to address cyber risks while avoiding new regulatory barriers without a congressional vote.
Show source details & analysis (3 sources)
📌 Key Facts
- On Tuesday, June 2, 2026, President Trump signed an executive order establishing a program for AI companies to voluntarily share powerful new models with the federal government before public release.
- The order asks companies to voluntarily submit their most powerful models for testing up to 30 days before public release — a shorter window than the 90 days contemplated in an earlier draft.
- The White House delayed a planned signing last month over concerns that the earlier 90-day review could stifle U.S. AI innovation and weaken American firms' position against China.
- The executive order directs federal agencies to develop benchmarks to assess AI models' cyber capabilities and to create an AI cybersecurity clearinghouse to review and share information about vulnerabilities.
- The order explicitly states that nothing in the voluntary testing framework authorizes creation of a mandatory government licensing, preclearance, or permitting requirement for developing or releasing new AI models, underscoring that binding regulation would have to come from Congress.
- Administration concerns were heightened after Anthropic said in April 2026 it limited release of its Mythos Preview model because of its ability to identify and exploit software security vulnerabilities, which helped spur the push for federal benchmarks and a cyber clearinghouse.
📰 Source Timeline (3)
Follow how coverage of this story developed over time
- On Tuesday, June 2, 2026, Trump signed an AI executive order that asks companies to voluntarily submit their most powerful models to the government for testing up to 30 days before public release, a shorter window than the 90 days contemplated in an earlier draft.
- The executive order directs federal agencies to develop benchmarks to assess AI models' cyber capabilities and to create an AI cybersecurity clearinghouse to review and share information about vulnerabilities.
- The order explicitly states that nothing in the voluntary testing framework authorizes creation of a mandatory government licensing, preclearance, or permitting requirement for developing or releasing new AI models, underscoring that binding regulation would have to come from Congress.
- The White House delayed a planned signing last month over concerns that the earlier draft order, including a 90-day review period, might stifle U.S. AI innovation and weaken American firms' position against China.
- The article notes that internal administration concerns were heightened after Anthropic announced in April 2026 that it was limiting release of its Mythos Preview model because of its ability to identify and exploit software security vulnerabilities, which helped spur the push for federal benchmarks and a cyber clearinghouse.
- On Tuesday, June 2, 2026, President Trump signed an executive order establishing a program for AI companies to voluntarily share powerful new models with the federal government before public release.
- CBS characterized the order in its on-air segment as giving the government an 'early look' at powerful AI models under a voluntary program.