The Trump administration has finalised a framework for vetting new AI models for safety and cybersecurity risks but plans to keep the policy private and share testing criteria only with select technology companies, according to The Guardian.
Key facts
- The White House finalised an AI testing framework this week after months of talks with industry leaders, according to The Guardian.
- The framework covers how new AI models will be tested for safety and cybersecurity risks.
- Staff from OpenAI, Anthropic, Meta, Google, Nvidia and Microsoft attended a private White House meeting on Tuesday, the report said.
- The vetting process is voluntary, and its details will not be released publicly.
- Testing criteria will be shared only with a select few technology companies.
After months of discussions with technology industry leaders, the Trump administration this week finalised a framework for how it will test new artificial intelligence models for safety and cybersecurity risks, according to The Guardian.
The White House is keeping the details of the framework private. The Guardian described the move as a blow to transparency.
On Tuesday, staff from OpenAI, Anthropic, Meta, Google, Nvidia and Microsoft attended a private meeting with White House officials to review the AI framework, according to the report.
Multiple outlets have since reported that while the voluntary vetting process for new AI models has been settled, the White House does not plan to release its policy publicly. Instead, it will share testing criteria with only a select few technology companies.
The Guardian said the framework leaves open questions about how the process will operate and be overseen. The report did not detail the framework’s specific testing criteria, and the White House has not made them public.
Why it matters
How governments test powerful AI systems for safety and security affects everyone who interacts with those tools. Keeping the vetting criteria private makes it harder for the public and independent experts to assess whether the process is rigorous or effective.
Frequently asked questions
Is the White House AI testing process mandatory?
No. According to The Guardian, the vetting process for new AI models is voluntary.
Which companies were involved in reviewing the framework?
Staff from OpenAI, Anthropic, Meta, Google, Nvidia and Microsoft attended a private White House meeting on Tuesday to review the AI framework.
Will the public be able to see the testing criteria?
No. The White House does not plan to release its policy publicly and will share testing criteria only with a select few tech companies.