Skip to main content
Policy & Ethics

OpenAI Introduces Watermarking for ChatGPT Outputs in the EU

OpenAI will automatically watermark ChatGPT outputs in the European Union, driven by regulatory requirements.

By Precis Daily Newsroom1 min read108 words
Illustration for: OpenAI Introduces Watermarking for ChatGPT Outputs in the EU
Illustration
Key points
  • Outside the EU, watermarking will be optional, but available to users.
  • OpenAI will automatically watermark ChatGPT outputs in the European Union, according to Ars Technica: AI.
  • The EU AI Act aims to promote transparency and accountability in AI-generated content.

What happened

OpenAI will automatically watermark ChatGPT outputs in the European Union, according to Ars Technica: AI. This move is driven by regulatory requirements in the region.

Why it matters

The EU AI Act aims to promote transparency and accountability in AI-generated content. The watermarking feature is designed to identify AI-generated content without significantly altering the quality of the output.

Key details

Outside the EU, watermarking will be optional, but available to users. The watermarking tool is based on OpenAI's proprietary method, called textGrain, which puts patterns in word choices that are not noticeable to human readers but can be detected by a specialized detector.

Sources

Summarized from the linked originals.

Related stories

Illustration for: Whatever AI Safety Is, It’s Not This
Policy & Ethics

Here’s a fun history lesson for you. Don’t worry, we’ll make it quick! In 1966, nearly 51,000 people died on US highways. The overwhelming evidence suggested that many lives could have been saved with modern seat belts , which by then had been around for several years.

WIRED AI4 min