Have 30 minutes? Then watch this video with a fantastic and clear explanation of the Watermarking in AI Generated content from Computerphile
Not only explains what watermarking is, but also has an explicit mathematical demonstration through an example of how it works.
And for the most of the users of state-of-the-art models, how to detect it? One needs to wait for the API from the providers, as the keys will not be public.
Anthropic has already signed the “EU AI Act’s Article 50(2) Code of Practice on Transparency of AI-Generated Content” and (for now) organizations need to register to get access to the Claude Watermark Detector Access
The quotes I retained?
- If you alter one thing, and another thing, and another, in an essay – why not writing it yourself?
- How easy is it to remove watermarking from Code generation? Possibly by removing / editing most of the comments.
Are you changing anything in your workflow due to the Watermark?
