←back to thread

SynthID – A tool to watermark and identify content generated through AI

(deepmind.google)
110 points jonbaer | 2 comments | 30 Aug 25 03:29 UTC | HN request time: 0.406s | source
Show context
teiferer ◴[30 Aug 25 06:27 UTC] No.45072366[source]▶
Could anybody explain how this isn't easily circumvented by using a competitor's model?

Also, if everything in the future has some touch of AI inside, for example cameras using AI to slightly improve the perceived picture quality, then "made with AI" won't be a categorization that anybody lifts an eyebrow about.

replies(5): >>45072388 #>>45072596 #>>45072693 #>>45073014 #>>45074013 #
1. QuadmasterXLII ◴[30 Aug 25 12:23 UTC] No.45074013[source]▶
I wonder if this will survive distillation. I vaguely recall that most open models answer “ I am chat gpt” when asked who they are, as they’re heavily trained on openai outputs. If the version of chatgpt used to generate the training data had a watermark, a sufficiently powerful function approximator would just learn the watermark.
replies(1): >>45074218 #
2. xpe ◴[30 Aug 25 12:53 UTC] No.45074218[source]▶
Are you expecting a distilled model to be sufficiently powerful to capture the watermark? I wouldn’t.

Additionally, I don’t think the watermark has to be deterministic.