US government limits export of powerful AI models to cybersecurity defenders
TechCrunch AI
The US government has restricted access to Anthropic’s Fable and Mythos AI models, citing national security concerns, leaving cybersecurity experts without the tools to find vulnerabilities and make software more secure. The move has sparked an open letter from 76 experts, including well-known figures, arguing that the action is ‘dangerous’ and could be used by adversaries. The models were initially restricted to prevent malicious hackers, but the new export controls have further limited their use.
💡 Key Takeaways
- The US government has restricted access to Anthropic's Fable and Mythos AI models, citing national security concerns, leaving cybersecurity experts without the tools to find vulnerabilities and make software more secure.
- The move has sparked an open letter from 76 experts, including well-known figures, arguing that the action is 'dangerous' and could be used by adversaries.
- The models were initially restricted to prevent malicious hackers, but the new export controls have further limited their use.