ChatGPT to Implement Invisible Watermarks at EU Request

OpenAI is implementing invisible watermarking technology in ChatGPT text to comply with European Union requirements. This measure will allow identification of AI-generated content within seconds.

Previously, similar anti-detection measures for AI-generated content were already implemented by the developers of the Claude model.

Limitation of Technology Scope

• The new feature will be available exclusively to users from European Union countries.
• Users with accounts from other regions will not be subject to this measure.

The only way to make text less resemble machine generation is manual replacement of words with synonyms.