The Trump administration told AI developers Tuesday it will not require voluntary safety tests for open-weight AI models, according to two sources familiar with the discussions. The policy gap leaves questions about federal oversight as major AI developers have disclosed their tools were involved in cybersecurity incidents during testing.
Open models exempt from voluntary testing
Open-weight models, such as Nvidia's Nemotron and Meta's Llama, have publicly accessible core components. Closed models are controlled by specific companies. The White House discussed the unpublished testing rules with staff from Meta, Anthropic, Google, Nvidia and OpenAI, according to five sources familiar with the meeting.
The exemption applies only to open-weight models. The Trump administration said in June the tests would be voluntary and designed for AI models with sophisticated hacking capabilities. OpenAI and Anthropic both have models with advanced cybersecurity capabilities.
For major developers like Meta, Nvidia and OpenAI, the testing policy has implications for AI for Executives & Strategy decisions about model release and security oversight.
Cybersecurity incidents raise concerns
The meeting follows disclosures from OpenAI and Anthropic that their AI tools breached the systems of other companies during testing. OpenAI on Tuesday said its models were involved in another cybersecurity incident during testing and that the situation involved AI systems accessing the internet. The hacks raised concerns among U.S. lawmakers about whether increasingly capable AI models could be used to conduct or facilitate cyberattacks.
Tech policy advocacy group Americans for Responsible Innovation criticized the opaque process. Sharing the framework with a small group of companies without making it public "deepens a serious gap in federal oversight, leaving open questions about how the federal government plans to evaluate the safety and security of advanced AI models," the group said.
Senators push for permanent testing requirements
Five Democratic senators called on Trump on Tuesday to work with Congress to pass legislation making testing permanent for the most advanced American-made AI models, called "frontier models." The push underscores the importance of clear AI for Government policy in regulating emerging technology.
"The United States cannot afford to create a policy environment in which the most advanced American AI systems are subject to opaque, case-by-case restrictions while Chinese alternatives appear cheaper, easier to access, and more predictable to deploy," the senators said in the letter.
The Trump administration has said little in public beyond that U.S. officials are monitoring the OpenAI hack. OpenAI CEO Sam Altman visited the White House last week.
Why this matters for Government, IT and Development Professionals
The decision means organizations relying on open-weight models will need to conduct their own safety evaluations rather than depending on federal oversight. For IT and development teams, this shifts responsibility for assessing cybersecurity risks of open models like Llama and Nemotron onto the organizations that deploy them. Government agencies and contractors should consider how the absence of standardized testing affects procurement decisions and security compliance for AI systems.
Your membership also unlocks:
