Anthropic expands text watermarking to older Claude versions beginning November 3
This isn't the first time Anthropic has issued a watermark, but the latest iteration will apply to Mythos 5, Opus 4.7, Opus 4.6, and Sonnet 4.6.
Anthropic has announced plans to expand text watermarking to older versions of its Claude AI, starting November 3. The move comes in response to EU legislators' calls for AI-generated text to include a hidden watermark, enabling differentiation between human and AI-generated content. Initially, the watermark was only applied to the most recent versions of Claude, but Anthropic has now extended it to include Claude Mythos 5, Claude Opus 4.7, Claude Opus 4.6, and Claude Sonnet 4.6.
Anthropic's watermarking technology involves an imperceptible pattern in word choice, ensuring it does not affect the accuracy, readability, or contain any identifying user information. This technology is the same as the one used in Opus 5.5. The AI giant maintains that the watermark aims to maintain the integrity of human-generated content in an era where AI-generated content is becoming increasingly prevalent.
The expansion of watermarking to older Claude versions highlights the growing concerns over the use of AI in various sectors, including legal, literary, and scientific research. As AI-generated content becomes more commonplace, there is a growing need to ensure that human input remains distinguishable. Lawyers, novelists, and researchers have all been caught using AI in their work, raising questions about the reliability and ethics of AI-generated output.
The introduction of watermarks may help address these concerns, but it remains to be seen whether they will effectively encourage humans to remain honest in their use of AI.
Written by urgent.news from Mashable's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.
