The U.S. Commerce Department ordered Anthropic to suspend exports of its advanced AI models, Mythos and Fable 5, due to concerns they could be diverted to military intelligence users in China and Russia. Anthropic has disabled global access to the models.
The U.S. government's actions highlight the growing concerns over the potential misuse of advanced AI technologies by adversarial nations and the complex balance between fostering innovation and ensuring national security.
The U.S. Commerce Department has ordered Anthropic to suspend the export of its advanced AI models, Mythos and Fable 5, citing concerns that they could be diverted to military intelligence users in China, Russia, or other countries of concern. Commerce Secretary Howard Lutnick sent a letter to Anthropic CEO Dario Amodei detailing the risks. In response, Anthropic stated it would disable access to the models globally.
The government believes there is a method to bypass safeguards that would prevent Fable 5 from being used in identifying software vulnerabilities, though Anthropic maintains that the bypass found only minor security flaws comparable to those in other publicly available models.
Senior Anthropic technical staff are scheduled to meet with Commerce Department officials to discuss the matter. This action follows earlier concerns about a potential diversion of the technology to a China-linked group. Additionally, over 80 cybersecurity executives and experts from major firms, including Nvidia and Adobe, signed an open letter supporting Anthropic and urging the administration to lift the restrictions.
Previously, Anthropic had warned about the hacking capabilities of its Mythos model and had held it back from wide release. The company rolled out a public version, Fable 5, on June 9, which included cybersecurity safeguards. Anthropic stated it works with the government to test models before release and received approval for Fable 5.