Anthropic Adds AI Watermark to Claude, Sparking User Cancellations

On August 15, artificial intelligence company Anthropic announced the addition of a watermarking feature to its AI text generator, Claude. The watermark indicates that the text has been processed by Claude, though Anthropic emphasized it cannot prove that Claude was the original author of the text. The watermark may remain even if Claude only performed editing tasks such as proofreading, translation, or summarization. As a result, some users have canceled their subscriptions, including public policy professionals, independent developers, and AI consultants, who worry the watermark could permanently tag their work, revealing AI usage and potentially triggering contractual penalties or compliance issues. Anthropic stated that the watermark was introduced to comply with the European Union’s Artificial Intelligence Act and plans to offer a free API for users to detect whether a piece of text contains a Claude watermark.