Anthropic has reversed a policy that restricted researchers from using its Claude Fable 5 model for competing AI development. The company admitted to making the wrong tradeoff by not disclosing that the model would reroute or degrade responses for tasks like training other LLMs. While the safeguards remain, Anthropic will now make these restrictions visible to users, apologizing for the lack of transparency and the negative impact on researchers. AI
IMPACT Ensures greater transparency in model capabilities, fostering trust and enabling more accurate research and development.
RANK_REASON Company reverses a policy that negatively impacted users and faced backlash for lack of transparency.
Read on Mastodon — mastodon.social →
AI-generated summary · Google Gemini · from 2 sources. How we write summaries →
