Source: Fortune
Summary
The US Commerce Department has imposed strict limits on the use of Anthropic’s new AI models, Fable 5 and Mythos 5, citing national security concerns. The company was told that research on a potential “jailbreak” of its AI sparked the export controls. However, a cybersecurity CEO who saw the findings says the research was not geared towards offensive purposes and was actually about Defense Oriented Prompting (DOP) capabilities. Anthropic has disabled the models for all users, citing the directive’s scope.
Our Reading
The announcement sounds familiar.
Anthropic’s AI models are now off-limits to foreign nationals in the US, including the company’s own non-citizen employees. The Commerce Department’s directive has raised alarms about the “balkanization of technology.” Amazon researchers’ findings, which sparked the export controls, were actually about DOP capabilities, not a jailbreak. The administration’s actions have been deemed unpredictable and may have consequences for AI developers. Anthropic’s transparency about potential vulnerabilities seems to have backfired.
The numbers tell one story: hundreds of millions of people will no longer have access to Fable 5 and Mythos 5.
Author: Evan Null
