Hacker Newsnew | past | comments | ask | show | jobs | submitlogin

Even the issue you mention, training without consent, won't stop this.

Chances are really very low that this will become illegal and enforceable. It would require some very draconian laws, whilst copyright legislation is low priority in government circles, even more so in these times.

Even if somehow this would be outlawed in the US, nobody cares internationally. Right now, on Amazon you can buy knockoffs of millions of products from China that violate IP/copyright. Nobody cares. Do you think they will care about something as worthless as a digital image? A digital image that can't even be reliably detected as being AI generated?

And there's yet another work-around. Scrape images that don't require consent or make consent part of terms and conditions. Google made Google Photos free for about a decade, and trained it for free on all your stuff.



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

Search: