Anthropic launched its Claude Fable 5 AI model on Tuesday. It is a “safer version” of the Mythos AI that has been making headlines for finding hidden critical software flaws on its own, potentially offering capabilities for misuse. Now, Microsoft seems to be hesitant about the company’s employees using this new AI model. According to The Verge, Microsoft is limiting its workforce from using Claude Fable 5 because of data retention rules.
Is Microsoft reviewing Claude Fable 5?
As per the publication, this move comes after Anthropic released update policies requiring certain user data to be stored for safety monitoring. The Satya Nadella-led firm is reportedly reviewing these changes through its legal teams to decide whether this AI model can be safely used internally.
The report notes that the issue majorly involves how customer data and confidential company information are likely to be stored and handled, especially when Claude Fable 5 is said to require temporary retention of prompts and outputs for up to 30 days.
Notably, the main concern for Microsoft is expected to be data privacy and compliance. The tech giant reportedly wants to ensure that sensitive data shared by employees and customers is not stored or misused.
What Are The Change In Anthropic’s Policies
Anthropic has reportedly brought a new data retention system for its ‘Mythos-class’ models, which now includes Claude Fable 5. As per this system, user prompts and AI-generated responses could be stored for up to 30 days for trust and safety monitoring.
Moreover, the report highlights that the Claude Fable 5 is not currently available in Microsoft’s internal version of GitHub Copilot, even though other Claude models are accessible. These previous AI models reportedly operate under stricter Zero Data Retention rules. It simply means that the previous AI models do not store user data after processing.
For those who are not aware, the Claude Fable 5 has been made available to the public. Anthropic claims that it is the “safer” version of the Claude Mythos Preview that was released earlier this year. The new AI model is now more powerful in performing tasks such as software development, data analysis and complex reasoning.
