OpenAI to watermark ChatGPT text in EU under AI Act
In short
OpenAI will watermark ChatGPT text in the EU under AI Act-related rules. The publisher summary says OpenAI joined the EU AI-content code with Google, Anthropic, Meta, Microsoft and Mistral.
Tests show the watermark can get much weaker after editing. The reporting does not say when the change starts or how strong the watermark will be in real use.
What the reporting says
OpenAI will watermark ChatGPT text in the EU under AI Act-related content code rules, according to the publisher summary. The report says OpenAI is a signatory to the EU AI-content code alongside Google, Anthropic, Meta, Microsoft and Mistral, and tests show the watermark can weaken sharply after editing. It is not stated in the available reporting when the change begins or how effective the watermark will be in practice.
Why it matters
OpenAI will watermark ChatGPT text in the EU under the AI Act-related content code, according to the publisher summary. The report also says OpenAI is a signatory alongside Google, Anthropic, Meta, Microsoft and Mistral, and that tests found the watermark can weaken sharply after editing. The rollout timing and real-world effectiveness are not stated in the available reporting.