Microsoft has restricted employee access to Anthropic's new Claude Fable 5 model due to concerns over its data retention policies. The model requires data retention for its safety classifiers, meaning Anthropic may store prompts and outputs for up to 30 days, or longer if flagged. Microsoft's legal teams are currently evaluating these changes, particularly regarding customer data and confidential information, to determine if internal use will be permitted. This restriction comes shortly after Anthropic's announcement of the Mythos-class model, which was initially deemed too dangerous for public release. AI
IMPACT Companies may need to re-evaluate their use of AI models with new data retention policies.
RANK_REASON A company is restricting the use of a third-party AI model due to policy concerns.
Read on Mastodon — mastodon.social →
AI-generated summary · Google Gemini · from 2 sources. How we write summaries →
