A leak reveals that Anthropic is testing a more capable AI model "Claude Mythos"
5 hours ago
- #Data Leak
- #Cybersecurity Risks
- #AI Development
- Anthropic is developing a new AI model called 'Claude Mythos' or 'Capybara', which is more capable than any previous model they have released.
- The model is currently being tested by early access customers and is described as a 'step change' in AI performance.
- Details of the model were leaked due to a 'human error' in the configuration of Anthropic's content management system, making draft blog posts and other materials publicly accessible.
- The new model poses significant cybersecurity risks, with capabilities far surpassing current models, potentially aiding hackers in large-scale cyberattacks.
- Anthropic plans a cautious rollout, focusing initially on cyber defenders to mitigate risks.
- The leak also revealed plans for an exclusive CEO summit in Europe, part of Anthropic's strategy to market its AI models to large corporate clients.
- Anthropic acknowledged the data leak and has since restricted public access to the compromised data store.