In brief: The US government receives 30 days of advance access to new powerful AI models to benefit from their vulnerability detection, while the tech industry was spared longer exclusivity periods.
President Donald Trump has signed an executive order granting US federal agencies exclusive access to powerful new AI models for up to 30 days before they are made available to the broader public under a planned voluntary framework agreement.
The original executive order provided for a 90-day exclusivity period for the US government, but was reduced to 30 days following resistance from the tech industry. Technology companies feared that longer advance access could disadvantage American firms in global competition, particularly against Chinese competitors. The signing had been postponed pending negotiation of a reduced version.
The regulation is being discussed in light of capabilities of new AI systems to identify profound security vulnerabilities. Anthropic’s Claude Opus 5 model (code: Mythos Preview), for example, discovers vulnerabilities in existing software that have remained undetected for decades. Such systems could enable the US government to harden its own infrastructure, but in hostile hands could also function as a cyber weapon and open up espionage potential. Anthropic distributes this model in limited form to enterprises and government agencies, without public release.
Security experts expect that AI systems will be massively instrumentalized by cybercriminals for attacks in the foreseeable future. Trump’s predecessor Joe Biden had introduced stricter requirements for security audits by AI providers; Trump lifted these on the grounds that they hindered the American AI industry.
Source: www.it-daily.net · Published 3 June 2026
Lumi AI News — AI-assisted curation in accordance with Art. 50 EU AI Act. Paraphrase and classification by Lumi News Pipeline v1.2.9.