Anthropic to Watermark All AI-Generated Text From Claude Models Starting August 2
Anthropic confirmed it will watermark AI-generated text and files from Claude models, complying with the EU AI Act's Transparency Code that took effect August 2. The watermark is applied at the model level and persists through copy-paste, according to the company.
Anthropic will watermark text generated by its Claude models to comply with the European Union's AI Act Transparency Code, which took effect on August 2, 2026. The company confirmed the move in an updated support page rather than a formal announcement.
The EU's Transparency Code requires AI companies to mark AI-generated or AI-edited content in a way that other systems can detect. Anthropic said all models released after August 2 will automatically include technology to watermark both computer-generated text and files. For files, the company said it is using the C2PA open standard, an industry specification already adopted by other AI and media companies for content provenance.
Anthropic also said it plans to extend watermarking support to older models, though it did not specify a timeline. According to the company's support page, the watermark is embedded directly in the generated text: "Because the watermark is part of the text, it will travel with the text when it's copied and pasted elsewhere, and may persist through some editing. Watermarking will be applied at the model level, which means it will be present no matter which Claude product or surface the text comes from."
The watermarking will apply across Anthropic's product line, including the Claude platform API, the Claude consumer app, Claude Code, Claude Cowork, and Claude Tag.
One open question: how much a user needs to edit AI-generated text before the watermark breaks or disappears. Anthropic has not disclosed the technical threshold, and TechCrunch has asked the company for clarification.
Anthropic joins a growing list of companies that have committed to the EU's Transparency Code, including Black Forest Labs, Google, Meta, Microsoft, OpenAI, and Synthesia. The push toward watermarking has accelerated across the industry in recent weeks. AI music platform Suno said last week it would mark tracks generated on its platform following legal challenges over copyright. Newsletter platform Substack partnered with detection firm Pangram last month to flag AI-generated content, with CEO Chris Best coining the term "Claudefishing" to describe writers passing off AI-generated text as their own.
What this means
This is a compliance move, not a voluntary transparency push. The EU AI Act's Transparency Code gives companies a legal deadline, and Anthropic's quiet support-page update — rather than a press release — suggests the company is treating this as a regulatory checkbox rather than a marketing opportunity.
The unresolved question about editing thresholds matters more than it might seem. If light edits strip the watermark, the system offers minimal real-world detection value against anyone motivated to evade it — students avoiding academic detection, or content farms avoiding platform bans. Watermarking that survives copy-paste but not paraphrasing catches casual users, not determined ones.
Expect this to become the industry baseline rather than a differentiator. With Google, Meta, Microsoft, and OpenAI already committed to the same EU code, watermarking is turning into table-stakes regulatory compliance for any AI company operating in Europe, not a competitive feature. The real test will be whether these watermarks hold up against third-party detection tools and adversarial editing in practice.
Related Articles
Anthropic Says New Defenses Cut Claude Reasoning-Token Extraction by 90–99%
Anthropic says new defenses have reduced the volume of Claude reasoning tokens extracted through distillation attacks by 90 to 99 percent. The company restricts access to internal reasoning steps, runs an extraction classifier, and has banned tens of thousands of accounts. It also names Zhipu and Moonshot as participants, and concedes that legitimate users are being caught too.
Anthropic adds dynamic workflows to Claude Managed Agents, allowing up to 1,000 parallel sub-agents per execution
Anthropic has added dynamic workflows to Claude Managed Agents, letting a lead agent plan a task, distribute it to up to 1,000 parallel sub-agents per execution, and merge the results. Anthropic claims the approach found 66 of 70 hidden bugs in a 116,000-line codebase, versus 14 to 27 for a single agent. Pricing and token costs were not disclosed.
Anthropic adds Dashboards and Motion betas to Claude; Docs, Slides and Design exit beta for all plans
Anthropic launched two beta features for Claude: Dashboards, which builds auto-updating dashboards from connected data sources, and Motion, which generates animated explainer videos exportable as MP4. Docs, Slides and Design leave beta and now work on every plan, including free accounts.
Google and Meta unveil frontier models they claim match OpenAI and Anthropic's best
Google and Meta each unveiled a new frontier model within a span of days, and both companies claim the models can compete with the best from OpenAI and Anthropic. Independent benchmarks, pricing and technical specifications were not available in the source reporting.
Comments
Loading...