startup · techcrunch.com

OpenAI will start watermarking ChatGPT’s text in the EU

OpenAI will start watermarking ChatGPT’s text in the EU
Image via techcrunch.com
What happened

On October 4, 2026, OpenAI announced that it will embed an invisible watermark in every piece of text generated by ChatGPT and Codex for users in the European Union to satisfy the EU AI Act. The watermark can be detected by OpenAI's verification tools but may become harder to spot after extensive user editing, creating a new layer of provenance tracking for AI‑generated content.

Why it matters

The move directly addresses regulatory pressure on AI providers to make generated content traceable, reducing the risk of undisclosed AI use in misinformation or academic cheating. It also sets a technical precedent that could force competitors to adopt similar provenance mechanisms, reshaping compliance costs across the industry.

Your next move — for a otherBe aware that any AI‑generated text you encounter from ChatGPT or Codex in the EU may carry an invisible watermark, so consider using verification tools before publishing or citing such content. If you edit the text heavily, run a detection check again to ensure the provenance signal remains intact for compliance audits.
Read full article ↗