OpenAI will begin automatically watermarking text generated with ChatGPT in the European Union, driven by a need to comply with the EU AI Act, and will offer the feature elsewhere but off by default.
OpenAI says it will give access to the detector to a limited number of researchers and organizations, with a request-for-approval process for others to be added over time.
OpenAI's tests of textGrain show a best-case 92 percent successful detection rate, but changing just 10 percent of the text reduces that rate by almost 30 percent, and changing 20 percent lowers it by almost 75 percent.
According to TechCrunch, it appears possible to edit out the watermark, so it remains to be seen whether it serves any significant purpose.
