Why Anthropic is watermarking text generated by Claude AI, how it will work
Anthropic, an AI company, is introducing watermarks for text generated using its Claude AI model. These watermarks will be imperceptible and will not alter the meaning, quality, or readability of Claude's responses. Additionally, Anthropic will attach signed provenance metadata to certain file types, such as images, to indicate whether they were processed using Claude and to prevent tampering.
This move is aimed at complying with the European Union's AI Act and a voluntary code of practice on transparency of AI-generated content. The watermarks will be embedded directly into the text and will persist even after editing. Anthropic will also develop tools to help users detect the presence of these watermarks and metadata.
However, watermarks have limitations, as they can be removed or edited in images and may be easily circumvented using AI models to paraphrase or rewrite text.
Written by urgent.news from The Indian Express's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.
