Busting Those Misleading Myths About Anthropic AI Watermarking During Proofreading Or Fixing Typos

InnovationAIBusting Those Misleading Myths About Anthropic AI Watermarking During Proofreading Or Fixing TyposByLance Eliot,Contributor.Forbes contributors publish independent expert analyses and insights. Dr. Lance B. Eliot is a world-renowned AI scientist and consultant.Follow AuthorAug 20, 2026, 03:00am EDTSummaryAnthropic's new AI text watermarking for Claude is widely misunderstood. The system embeds hidden watermarks by having the AI subtly alter word choices during text generation. This makes watermarks hard for humans to detect but identifiable by Anthropic authorized tools. Watermarks persist through copying but can potentially be broken by editing, though longer texts tend to be more resilient. Importantly, simply proofreading text without allowing AI changes does not watermark the original content; only the AI's generated word responses are said-to-be marked. Fixing typos without changing the intended word avoids watermarking, but allowing substitutions can embed them.Show More Clearing up the prevailing confusion of Anthropic's new watermarking of Claude AI-generated output.gettyIn today’s column, I examine and bust or straighten out various misleading myths regarding the recently released AI text-oriented watermarking feature of Anthropic. The mainstream news and social media have been making zany and incorrect claims about what this particular technique of watermarking is and does. This, in turn, has tended to create widespread confusion and undue consternation among people who are unsure whether their text will be watermarked or not when using Claude.Weighty questions on people’s minds include whether text that they submit to the AI for proofreading will end up watermarked, and whether the AI fixing incidental typos will also encompass the implanting of a watermark. To answer those questions, I will briefly lay out how it is that the watermarking actually occurs and explain how the hidden watermarks are infused into text (for my in-dept...