Anthropic's Claude AI model is reportedly implementing a form of invisible watermarking on its outputs, a feature that is not yet accessible or verifiable by users. This watermarking aims to identify AI-generated content, but its current inaccessibility raises concerns about transparency and user control. The development was noted in a discussion on Mastodon, highlighting a gap between the technology's implementation and its usability. AI
IMPACT Raises questions about transparency and user control over AI-generated content, potentially impacting content provenance and trust.
RANK_REASON Article discusses a feature of a model release but does not contain primary source information from the developer.
Read on Mastodon — mastodon.social →
AI-generated summary · Google Gemini · from 1 sources. How we write summaries →