OpenAI will start watermarking ChatGPT’s text in the EU
Published
What's this about?
OpenAI will watermark ChatGPT and Codex text in the EU to comply with the AI Act. Editing can make the invisible marks harder to detect, it says.
The one-panel cartoon
Read the cartoon's script
A bustling European newsroom filled with reporters and journalists poring over printed articles and documents, with a magnifying glass and watermarks barely visible on some pages.
Caption: Trying to spot the difference between journalism and watermarked text
Read the full 3-panel strip
Based on the original story on techcrunch.com.
Editor's note
This move by OpenAI shows a commitment to transparency and accountability in the digital age, helping users trace the origin of AI-generated content. While it's important to see how this plays out in practice, especially considering the ease with which watermarks might be removed, it's a positive step towards responsible AI usage.
Rate this comic
Read the script
Panel 1
Living Room *Characters: Alice (using a laptop), Bob (looking over her shoulder)* *Alice* (reading on the screen): *Bob* (curious):
Alice: Hey, OpenAI is going to start watermarking ChatGPT text in the EU!
Bob: Really? Why would they do that?
Panel 2
Living Room - Close-up of Alice's Laptop *Alice* (pointing at the screen): *Bob* (scratching head):
Alice: It's to comply with the new AI Act, so they can track where their content goes.
Bob: Interesting. So, it’s like a digital signature?
Panel 3
Living Room - Alice and Bob laughing* *Alice* (smiling): *Bob* (grinning): *Alice and Bob* (laughing together):
Alice: Guess they’re just being proactive. But I bet some tech-savvy user will find a way to remove it!
Bob: Probably, and then the cat-and-mouse game starts!
Alice: Here’s to the next batch of digital sleuths!
Comments