In a significant policy shift announced this Tuesday, the United States government has revealed that it will now have access to new artificial intelligence (AI) models from major tech giants to evaluate them before they are made available to the general public.
The move involves agreements with industry leaders including Google DeepMind, Microsoft, and xAI. This marks a departure from the Trump administration's earlier hands-off approach to regulation, which occurred as Silicon Valley rapidly deployed AI technologies that have fundamentally altered modern life.
According to officials, these partnerships are based on frameworks originally established during the Biden administration, which have since been renegotiated under the current leadership of Donald Trump. This development follows reports from the New York Times on Monday that the White House is considering an executive order to create a working group of government officials and tech executives to refine these review procedures.
Central to this initiative is the Center for AI Standards and Innovation (CAISI), an arm of the Commerce Department. CAISI, which replaced the US Artificial Intelligence Safety Institute created in 2023, stated that it will conduct targeted research and pre-deployment evaluations to better understand frontier AI capabilities and enhance overall AI security.
The immediate driver for this shift appears to be the emergence of "Mythos," a powerful new AI model developed by San Francisco-based start-up Anthropic. The company has refrained from releasing Mythos publicly, warning that its ability to identify software security vulnerabilities could lead to a major cybersecurity reckoning. Reports indicate that the National Security Agency has already gained access to Mythos for rigorous testing.
Highlighting the importance of the move, CAISI Director Chris Fall stated on Tuesday that "independent, rigorous measurement science is essential to understanding frontier AI and its national security implications."
Reference(s):
cgtn.com
