Anthropic introduces imperceptible watermarking for AI-generated text in Claude models

Published Aug 11, 04:09 AM UTC

Editorial responsibility

Prepared by: WorldAttention Editorial Systems
Human review: No named human review is recorded for this page.
Automated steps: Source reporting is collected, normalized, translated or condensed automatically when needed.
Page status: Automatically published source-backed update

How automation and AI are used · Report a correction or rights issue

Anthropic announced that new Claude AI models will embed an imperceptible watermark into AI-generated text, as part of its transparency commitments under the European Union AI Act. The watermark does not alter meaning or readability but persists through copying and some editing. Claude models launched on or after August 2 will support this feature, with plans to extend it to older models and provide third-party detection tools. The move impacts the publishing industry, which has faced controversies over AI-generated content, such as the withdrawn novel 'Call Me, I'll Hide the Body' and Hachette's 'Shy Girl'. However, Anthropic noted that heavy editing, paraphrasing, translation, or mixing with other writing can make the watermark undetectable. Google DeepMind previously introduced similar text watermarking in 2024 via its SynthID technology.

Source

All Content from Business InsiderWestern

Original headline: Anthropic just rolled out a tool that'll decimate some people's dreams of writing AI novels undetected

Original language: EN

Part of this Story

Anthropic Adds Invisible Watermarks to Claude AI Text Under EU AI Act

See the full Story: all sources and camp coverage

People and interests

Anthropic European Union Google DeepMind