Anthropic Watermarks Claude Text Under EU AI Act
Get the Tech newsletter
Daily tech — startups, AI labs, chips, the launches that shape the next decade. Free.
- Anthropic confirmed it will watermark text generated by its AI models, including Claude, to comply with European regulations, per an updated support page.
- The EU AI Act's Transparency Code, which took effect August 2, requires AI companies to mark AI-generated or edited content in a way other systems can identify them.
- All models released after August 2 will automatically include watermarking tech for both text and files, with Anthropic using the C2PA open standard for files and extending support to older models over time.
- The watermark is embedded within the text itself, so it travels when users copy and paste and may persist through some editing — though Anthropic declined to specify how much editing strips it.
- Claude products — including the platform API, Claude, Claude Code, Claude Cowork, and Claude Tag — will all carry the watermark at the model level.
- Black Forest Labs, Google, Meta, Microsoft, OpenAI, and Synthesia have also committed to adhering to the EU code, joining a broader wave that includes Suno for music and Substack (partnering with Pangram) for newsletters.
Why it matters: Anthropic becomes one of the first major model providers to implement watermarking across its full Claude product line under a regulation that took effect August 2, giving platforms and regulators a traceable fingerprint on AI text that survives copy-paste downstream — though Anthropic has not disclosed how much editing is needed to strip it, leaving the durability of the mark an open question for users and compliance teams.
Ask SkimNews
