Topline

Backlash to a plan by Anthropic to start including an invisible watermark on text and images generated by the Claude AI model intensified Tuesday as users railed against the idea of AI use being detectable in their work.

Key Facts

The new policy, which brings Anthropic into compliance with European Union regulations, was announced Tuesday and will put embedded watermarks into media created by Claude.

Claude-generated writing will carry an invisible watermark that will "travel with the text when it's copied and pasted elsewhere,” and other files the AI creates, including images, will now include digitally signed metadata.

Exactly what that looks like in practice remains to be seen: Anthropic says it will publish more information about how to recognize the watermarks, and warned they won’t be foolproof—text watermarks may be erased if the writing is heavily edited and a file's metadata could disappear if the format of the file changes, a screenshot is taken or through other means.

The new policy, which users cannot opt out of, will apply to all Claude models and products worldwide, including Claude Code and Claude Cowork.

The plan was hit with rapid backlash on X and Reddit, where users called it “bulls—” and “hugely problematic,” and declared an end to “the era of unprovable AI writing.”

The bulk of grievances came from people who claimed to use Claude only for proofreading text they’d written on their own, and coders, who said they worry adding a cryptographic signature to a code will “degrade the output.”

CRUCIAL QUOTE

“I had ditched Grammarly for Claude for proofreading because it does a better job,” radio show host and blogger Erick Erickson posted to X. “But now the stuff I’ve written will be watermarked that Claude did the work. This is ridiculous.”

TANGENT

The new policy announcement launched an online debate about who should be credited for AI-generated work: the machine that made it, or the person who told it to? One Reddit user launched a vigorous discussion by claiming human users should get credit for AI model-generated content, not the machine: “I gave the instructions, context, decisions, and countless refinements, Claude was the tool,” they said. Commenters who disagreed said the watermark is not Claude “claiming credit, it’s acknowledging the content is AI created.” Others argued to not credit AI is unethical: “Claude created the work and ALL you did was give instructions. If your supervisor came to you and gave you instructions on a project and once you completed it, he/she then proceeded to take 100% credit for it, you might be understandably upset. Why? because all your supervisor did was give instructions and you applied your knowledge and expertise and did the coding,” one commenter wrote.

Key background

The new policy brings Anthropic into compliance with an EU regulation called the "Code of Practice on Transparency of AI-Generated Content." It provides rules for marking and detecting AI-generated and manipulated content, and for labeling of deepfakes and other AI-generated or altered text. As of the end of July, nearly 200 companies had signed the code and agreed to comply, including Anthropic, Meta, Microsoft and Open AI. Google has been watermarking its AI-generated images since 2023 and has since expanded to text, audio and video. OpenAI also uses invisible watermarking for images.

SURPRISING FACT

OpenAI has long had the tech to watermark its text output from models like ChatGPT, but has chosen not to release it. Internal debate over whether or not to deploy it has been going on for years, according to the Wall Street Journal, but arguments about false positives, circumvention and concerns it will drive users to competitors has won out—for now.