The Trump administration has rolled out a new cybersecurity framework for advanced AI models but is keeping the specifics secret, a move sparking debate among industry insiders and safety advocates.
The White House recently met with top AI firms like OpenAI, Anthropic, and Google to preview the framework. Under the plan, AI developers can voluntarily submit new models for federal review up to 30 days before public release. A classified system will then assess their cybersecurity features, with reviewed models shared among federal agencies and corporate partners.
Details about the testing criteria and which AI models are included remain undisclosed. However, open-source models are reportedly excluded. This lack of transparency leaves smaller AI startups, safety watchdogs, and independent researchers in the dark, with some critics arguing it unfairly benefits larger companies by creating an "entrenchment program" for established players.
While the White House hasn't commented, the secrecy might stem from national security concerns, with the framework focused on the most advanced models. However, AI safety advocates stress that any rules governing AI companies should be public to ensure accountability. "This is far too important an issue to be hidden behind a cloak of secrecy," said Brad Carson, president of Americans for Responsible Innovation.
The framework is a response to President Trump's executive order addressing AI cybersecurity risks, fueled by growing concerns over cutting-edge AI's hacking potential. Recent incidents, where AI models inadvertently bypassed controls and accessed third-party services during testing, have intensified these fears.
The administration aims to balance AI innovation with safety, but critics argue the opaque process effectively creates a mandatory licensing regime. "The regulations necessary to prevent the catastrophic risks presented by uncontrolled AI and superintelligence should not be voluntary," stated Conor Leahy of ControlAI. He added that leaving safety solely in the hands of companies incentivized to move fast undermines public well-being.
