White House orders Anthropic to shut down new AI model within 90 minutes
The administration directed Anthropic to deactivate its recently launched Fable model after security concerns, giving the company a brief window to comply.
The White House instructed Anthropic to take its newly released Fable AI model offline, citing concerns that the system's safeguards against hacking could be circumvented. The directive was delivered three days after the model’s launch and allowed the company only 90 minutes to comply. Two sources familiar with the discussions, who asked to remain anonymous, said officials provided limited details about the specific risks. The request marks a notable step in federal oversight of advanced artificial-intelligence systems.
"We are taking this action to protect national security and public safety," a White House spokesperson said, according to the sources.
Anthropic has not publicly commented on the order, and the administration has not released further information about the alleged vulnerabilities.
