OpenAI is adding text watermarking in ChatGPT and Codex
By Stevie Bonifield · Oct 5, 2026, 1:08 PM CDT
An invisible, machine-readable watermark in text output is rolling out to ChatGPT and Codex, but only for users in the European Union at first. OpenAI says its textGrain watermarking "matched or exceeded" other approaches like Google DeepMind's SynthID for text, which is also the basis for the watermarking Anthropic announced in August . Like OpenAI, Anthropic made the move to meet the requirement
Excerpt shown under fair-use limits. Full text remains with the original publisher.
People in this coverage
Explore their history and attributable record. Being mentioned does not imply endorsement.
Layer 1 · Claims & fact checks
AI analysisLayer 3 · Reporting analysis
AI analysisAppeal to AuthorityThe article cites well‑known AI entities to bolster the credibility of OpenAI's watermark without providing independent evaluation.
Compliance FramingPositioning the rollout as a response to regulation frames the change as responsible and necessary.
Positive SpinThe mention of benchmark scores emphasizes that the watermark does not degrade performance, presenting the feature in a favorable light.
Context
AI analysisMissing context
The article does not explain how the watermark works technically, how it will affect user experience, any privacy implications, or how non‑EU users will be treated. It also lacks perspectives from independent experts or regulators on the effectiveness of the watermark.
Important context
The rollout is tied to compliance with the European Union's AI Act, which imposes transparency requirements on AI systems deployed in the EU.
Opinion vs. reporting
AI analysisThe piece is primarily reporting factual announcements but includes promotional language quoting OpenAI’s self‑assessment without independent verification, blurring the line between pure reporting and marketing.