Anthropic to Apply Invisible Watermarks to AI-Generated Content

AI company Anthropic has announced it will begin embedding machine-readable watermarks into content produced by its Claude models. This initiative aims to comply with new European Union regulations requiring greater transparency in artificial intelligence.

The watermarks will be invisible to human users but will allow for easier detection of AI-generated material by individuals and online platforms. Anthropic stated that text will carry embedded watermarks, and image files will include digitally signed provenance metadata where supported, facilitating content traceability.

These updates represent a future commitment rather than an immediate rollout. They are designed to meet anticipated regulatory obligations, including those mandated by the EU's AI Act, which seeks to ensure that AI-generated content is identifiable.

The introduction of these invisible markers is intended to help combat the spread of misinformation and enhance trust in digital content. This move positions Anthropic to address growing concerns about the authenticity and origin of information in the digital age.