Anthropic, which claimed AI model was too risky for public to use, releases ‘safe’ version
Anthropic has announced the release of a new version of its Mythos model, which is equipped with safety features to make it suitable for the general public. This move comes after the company raised concerns about potential safety threats posed by its latest AI model.
The new version, called Claude Fable 5, includes guardrails around cybersecurity and biological weapons. If users ask the model about off-limits topics such as exploiting software bugs or creating bioweapons, the response will be blocked, redirecting them to an older model known as Opus 4.8, according to Anthropic’s press release.
Anthropic, under the leadership of Chief Executive Dario Amodei, has long emphasized the potential dangers of its products while positioning itself as a responsible guardian of powerful AI technology. The San Francisco-based company is set to go public this year, competing with rivals OpenAI and Elon Musk’s SpaceX, which merged with his AI venture xAI earlier this year.
According to Anthropic, the release of a highly capable model like Fable 5 comes with risks. To mitigate these risks, the model has been equipped with safeguards that redirect queries on sensitive topics to the Opus 4.8 model, the next-most-capable in the lineup.
While some harmless requests may occasionally be blocked, Anthropic assures that such occurrences will be rare.
Anthropic claimed model public Releases risky safe Version
