United States centralizes control over frontier AI access
Tap title to read story
The U.S. government has assumed a decisive role in determining which companies may access cutting‑edge cybersecurity-focused AI models from Anthropic and OpenAI, according to a CNBC report on Friday citing people familiar with the matter. Previously, the labs independently managed access through Anthropic’s Project Glasswing and OpenAI’s Daybreak initiatives. The White House recently blocked Anthropic’s Claude Mythos 5 and Fable 5 over national security concerns, restoring access only after weeks of negotiations. This week it launched Gold Eagle, an AI clearinghouse for cyber vulnerabilities that will effectively make the administration the gatekeeper for new model access amid intensifying competition from China’s Moonshot AI.
More context
Timeline
- Last month, White House blocks Anthropic models
- Last month, negotiations restore Anthropic model access
- June 2024, OpenAI restricts models to trusted partners
- This week, White House launches Gold Eagle clearinghouse
- This week, Gold Eagle centralizes frontier AI approvals
- Same day, Moonshot AI releases Kimi K3 model
- Same day, Kimi K3 matches leading benchmark performance
- Friday, CNBC reports White House access shift
Why it matters
The White House's new role in AI access could affect your online safety. It's now the gatekeeper for cutting-edge cybersecurity AI. This means it decides which companies can use these tools to protect your data. Check if your service providers are on the approved list.
JQJO Editorial Desk · Human reviewed · AI assisted. Report a correction.
