OpenAI will watermark ChatGPT outputs by default—but only in the EU
Summary
OpenAI will start adding a hidden watermark to text created by ChatGPT in the European Union to follow the new EU AI law. This watermark helps detect AI-generated text but is not fully reliable and can be turned off outside the EU.Key Facts
- OpenAI’s watermarking feature is automatically turned on for ChatGPT texts in the EU.
- The EU AI Act requires AI-generated content to be marked so tools can identify it.
- The watermark uses a secret pattern in word choices that does not affect how the text reads.
- Detection works about 92% of the time but drops if the text is slightly changed.
- The watermark will be optional outside the EU and in OpenAI’s API service.
- OpenAI will give only a few trusted groups access to the watermark detector.
- Another AI company, Anthropic, has a similar watermarking system enabled worldwide.
- The watermarking feature for EU users will be available in the coming weeks.
Read the Full Article
This is a fact-based summary from The Actual News. Click below to read the complete story directly from the original source.
