Codex

Binance AI

Light

<p><strong>OpenAI</strong> will embed an invisible watermark in ChatGPT and Codex text in the European Union, though its own tests show swapping 10% of words cuts detection to 66%.</p>
<blockquote>
 <h2 id="key-points">Key Points:</h2>
 <ul>
  <li>The watermark reaches eligible ChatGPT and Codex users on all plans in the EU over the coming weeks and stays opt-in for API customers.</li>
  <li>Replacing 25% of words with synonyms cut detection to 17% in the company's own evaluation.</li>
  <li>Detector access is limited to approved researchers and expert organizations at launch.</li>
 </ul>
</blockquote>
<h2 id="openai-watermark-rollout">OpenAI Watermark Rollout</h2>
<p>The company <a href="https://openai.com/index/eu-text-provenance/" rel="noopener noreferrer nofollow">announced</a> the plan Monday in a blog post, saying the marking will reach eligible ChatGPT and Codex users on all plans in the EU over the coming weeks. It will not become a global default at launch.</p>
<p>Developers using the company's API anywhere in the world can <a href="https://techcrunch.com/2026/10/05/openai-will-start-watermarking-chatgpts-text-in-the-eu/" rel="noopener noreferrer nofollow">turn</a> it on for select models starting now, though it stays off by default. The move answers the EU AI Act, whose marking obligations took effect Aug. 2 and require providers to make AI-generated text identifiable by machines. Companies already on the market <a href="https://www.engadget.com/2277866/openai-will-add-a-digital-watermark-to-text-and-code-generated-in-the-eu/" rel="noopener noreferrer nofollow">have</a> until Dec. 2 to comply.</p>
<p><strong>Also Read:</strong><a href="https://yellow.com/news/binance-ai-agent-strategies-sentence-usdc" rel="noopener noreferrer nofollow">Binance AI Agent Builds Strategies From Plain Words, Charges 19.99 USDC To Trade</a></p>
<h2 id="textgrain-detection-limits">TextGrain Detection Limits</h2>
<p>The method, called textGrain, adds no visible symbol but subtly shapes the model's word choices, leaving a statistical pattern that a detector can pick up and that travels with copied text. OpenAI said it matched or beat other approaches it tested, including <strong>Google</strong>'s SynthID for text, and that it plans to release the technology as open source.</p>
<p>Editing weakens that signal.</p>
<p>In an evaluation of 400-token passages, replacing 10% of words with synonyms reduced detection from about 92% to 66%. Replacing 25% of words <a href="https://thenextweb.com/news/openai-text-watermarking-chatgpt-codex-eu-ai-act" rel="noopener noreferrer nofollow">cut</a> it to 17%. Short passages, math answers and translated text are also harder to flag, with the detector catching about 80% of 200-token passages on topics such as psychology at a 1% false positive rate.</p>
<h2 id="chatgpt-detector-access">ChatGPT Detector Access</h2>
<p>OpenAI is limiting its detector to approved researchers and expert organizations for now, citing the risk of missed watermarks and false positives. It also cautioned that a missing watermark "does not prove human authorship," and said the mark does not identify the user, the account or the prompt behind a passage.</p>
<p>The step follows <strong>Anthropic</strong>, which said in August that it would watermark text generated by Claude worldwide, a decision that drew complaints from some users. OpenAI had built a text watermark before but held it back, partly over concerns that users would switch to rivals, according to a 2024 news report. Both companies, along with Google, <strong>Meta</strong> and <strong>Microsoft</strong>, have <a href="https://digital-strategy.ec.europa.eu/en/news/strong-backing-code-practice-transparency-ai-generated-content" rel="noopener noreferrer nofollow">committed</a> to the EU's code of practice on AI-generated content.</p>
<p><strong>Read Next:</strong><a href="https://yellow.com/news/meloni-trademark-voice-stop-ai-deepfakes" rel="noopener noreferrer nofollow">Giorgia Meloni Wants A Trademark On Her Voice, Will It Stop AI Deepfakes?</a></p>

ChatGPT Watermark Heads To EU With A Catch: Light Edits Cut Detection To 66%

Table of Contents

Yellow News

Other articles published on Oct 6, 2026

Join our free newsletter for daily crypto updates!