Anthropic planning to release Mythos-capable models to the public

In a blog article, Anthropic has announced its intention to eventually release models that match the performance of its Claude Mythos Preview AI model to the public, once adequate safety measures are in place. Introduced in April, Mythos is deemed highly proficient at identifying software vulnerabilities. Nevertheless, due to the risk of cybercriminals exploiting undiscovered flaws, Anthropic restricted access to select partners through its Project Glasswing initiative.

In an update on Project Glasswing, Anthropic reveals plans to expand the initiative further, including with the US and allied governments, as well as other critical partners. However, the firm admitted that no AI company has yet developed safeguards robust enough to prevent misuse. 

Training Announcement: The BCS Foundation Certificate in AI examines the challenges and risks associated with AI projects, such as those related to privacy, transparency and potential biases in algorithms that could lead to unintended consequences. Explore the role of data, effective risk management strategies, compliance requirements, and ongoing governance of the AI lifecycle and become a certified AI Governance professional. Find out more.