Hacker News new | past | comments | ask | show | jobs | submit login

Possibly, but should be easy enough to exclude.

For example Dall-e has the pixels watermark (bottom-right), and I assume there's a possibility of an indicator that may by hidden in the data itself. One could also exclude common meme formats and their derivatives. Then there's the option of mapping to produced content via hashing a la Shazam, or have a discriminator component etc.

But you're right, it's not trivial. I just don't think it's too big of a deal.




You're assuming a _voluntary_ action _can_ be done. This won't happen, Lazy humans will feed non-aware AIs.

The OP is asking if s.o has "already thought of such an issue"




Join us for AI Startup School this June 16-17 in San Francisco!

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: