Anthropic pulls two AI models worldwide after US request

Amazon CEO Andy Jassy's safety concerns triggered Anthropic to shut down two AI models globally. According to Wall Street Journal, Amazon researchers found that Claude Fable 5 could be used to access information for cyberattacks. The US government then imposed export controls on Fable 5 and Mythos 5, leading Anthropic to disable them worldwide this Friday.

Amazon, a key investor in Anthropic, had previously raised safety concerns. David Sacks, former US AI chief, said a trusted partner found a 'jailbreak' vulnerability in the models. The US asked Anthropic CEO Dario Amodei to fix it or remove the models, but he refused. Anthropic argued that similar features exist in other public AI models.