Senate Intelligence Bill Would Encourage Prerelease Testing of AI Hacking Capabilities (May 26, 2026) - VitalLaw.com

The Senate Intelligence Committee passed legislation last week by a vote of 14-3 that would permit U.S. intelligence agencies to conduct prerelease testing of AI models capable of hacking critical infrastructure. The bill, titled the Senate Intelligence Bill, aims to encourage such testing before models are released to the public. The committee's vote occurred on May 26, 2026, according to the report.

The legislation would give intelligence agencies the opportunity to evaluate AI models for their potential to compromise critical infrastructure systems. The testing would take place prior to the public release of the models. The bill passed with broad support in the committee, with 14 votes in favor and 3 against.

The measure reflects ongoing congressional efforts to address risks associated with AI capabilities. The bill now advances to the full Senate for consideration. Further details on the bill's provisions and next steps have not yet been disclosed.