Anthropic

"this powerful 'Mythos-class' model has the potential to do profound good for the world"

"may sometimes catch harmless requests"

"Congress should demand AI firms conduct independent safety evaluations for their most advanced models."

"We are complying with the government’s legal directive and are removing access to Fable 5 and Mythos 5 for all users."

"To date, the government has only given us verbal evidence of a potential narrow, non-universal jailbreak."

"The net effect of this order is that we must abruptly disable Fable 5 and Mythos 5 for all our customers to ensure compliance."

""These campaigns are growing in intensity and sophistication. The window to act is narrow, and the threat extends beyond any single company or region.""