Anthropic launches an API that allows regulators, media and researchers to verify whether a text contains Claude's digital labeling.
The change comes against the backdrop of the EU AI Act's requirements for invisible labels in AI-generated text. The technology affects not only model developers but also organizations that need to assess the origin of materials.
Critics warn that such labeling could degrade text quality. A separate issue is transparency in situations where contracts prohibit AI use: the source does not specify how this conflict will be resolved.
editorial commentary
Why it matters
A likely consequence is increased external control over the origin of AI text and new requirements for editorial offices, researchers and regulators. The next observable signal will be publication of access terms, technical metrics, and independent assessments of API accuracy. Substantial uncertainty remains around the impact of labeling on text quality and contractual restrictions on AI use.