Anthropic Implements AI Watermarking to Meet EU Regulations

Anthropic has begun watermarking text generated by its Claude AI models, according to reporting from TechCrunch. The company is inserting invisible code into the chatbot's editorial outputs to comply with the EU AI Act’s Transparency Code, which mandates that tech firms label AI-generated or AI-edited content in a way that is identifiable to computer systems.

While the move satisfies European regulators, it has triggered significant pushback from some users on Reddit. One critic argued that the watermarks are "unethical" and "disgusting," contending that the human user performs the majority of the labor by providing the context, instructions, and refinements. This user questioned what credit the AI is claiming if the chatbot is merely a tool used to facilitate arduous work.

Another user, identified as visionode, described the system as a "draconian conspiracy" that will unfairly target average users—such as students reorganizing paragraphs or writers seeking synonyms—with a "digital tattoo." Other critics pointed to a perceived hypocrisy, noting the irony of watermarking outputs when frontier models were trained by scraping existing data.

However, TechCrunch reports that many other users support the policy. Some argue there is no valid reason to oppose the system other than a desire to deceive others, while others maintain that detecting AI-generated content is necessary to mitigate various risks.