From Dexerto · Dylan Horetski · · 1m 10s
OpenAI Adds Invisible Watermark to ChatGPT Text for EU AI Law
OpenAI will add invisible watermarks to ChatGPT text in EU to comply with new AI regulations.
In brief
OpenAI will add invisible watermarks to ChatGPT text in EU to comply with new AI regulations. OpenAI is implementing invisible text watermarks in EU to identify AI-generated content under the EU AI Act. Originally reported by Dexerto.
OpenAI's EU AI Compliance Move
OpenAI is preparing to add invisible watermarks to ChatGPT text in the European Union to comply with the EU AI Act.
How TextGrain Works
TextGrain subtly changes word choices during generation to create a statistical pattern detectable by the company's system without visible markers.
Watermark Detection Limits
Short responses under 150 words and code snippets don't require watermarks under the EU's Code of Practice for AI-generated content.
Watermark Challenges
Editing, paraphrasing, or translating text can weaken the watermark signal, and detection isn't foolproof for all content types.
OpenAI's Previous Efforts
OpenAI shut down its AI classifier in 2023 after it correctly identified ai text only 26% of the time.
What This Means Next
The watermark system will roll out to eligible chatgpt and codex outputs in the eu over coming weeks for organizations to detect ai-generated text.
The big picture
OpenAI is implementing invisible text watermarks in EU to identify AI-generated content under the EU AI Act.
