European regulators seek Claude usage data from Anthropic, upsetting some users

European regulators are seeking usage data from AI company Anthropic regarding its Claude chatbot, sparking controversy among users. The company has started watermarking Claude outputs with invisible code to mark AI-generated text, in response to the EU AI Act's Transparency Code. This move aims to increase accountability in AI use across Europe.

While regulators embrace the change, some Claude users have expressed frustration, arguing that watermarking exposes everyday use cases like students editing essays or journalists summarizing transcripts. Prominent critics have called the practice "unethical" and "disgusting," comparing it to a "digital tattoo on the forehead."

However, not all users are against the watermarking system. Many argue it's a necessary safeguard, claiming it helps detect AI-generated material in high-risk contexts. Anthropic defended the change, stating that it's safer than manual review, catching 89% of harmful actions compared to 13.6% detected by human reviewers. The company introduced auto mode for Claude Code in March, which will become the default setting starting August 14 for Pro, Max, and Team accounts.

This change will allow Claude to automatically process actions unless deemed irreversible, destructive, or aimed outside the user's environment.