5 ms·
This is just legal cover until such time as its possible to enforce no child exploitation imagery, no copyright stuff, etc. It doesn't matter if they are able
by rngname22 3y ago
This is just legal cover until such time as its possible to enforce no child exploitation imagery, no copyright stuff, etc.
It doesn't matter if they are able to enforce it, Valve can use this policy as cover if they ever get sued.
Don't overthink the motivation. They will not even have a bulletproof way to detect AI imagery as it evolves every single day as an arm's race and detection is a full-time job. Even a FAANG or a state actor would need to dedicate team(s) to detection technology and still have false negatives.
The same sorts of things already happen for example on YouTube and Twitch, where types of content are against TOS or copyright but enforcement is sporadic and selective, smaller operations often fly under the radar of enforcement, bigger creators who are netting the org sufficient revenue will likely be able to get away with more, etc, the automated tools for detection are flawed.
- birdyrooster 3y agoGoing public about your awareness to a problem necessitates an enforcement response to be considerable. Imagine you are a trademark holder and someone is using your IP but you don't enforce your trademark by litigating. Your claim is weakened. It shows the public and the court how significant this problem is for your party. Edit: copyright -> trademark
- epakai 3y agoYou seem to be confusing copyright and trademark. Copyright isn't diminished by non-enforcement. A trademark risks being invalidated or genericized when not enforced. Intellectual property is an encompassing term that seems to lead to this sort of confusion.
- birdyrooster 3y agoThank you!!
- MetaWhirledPeas 3y ago> They will not even have a bulletproof way to detect AI imagery as it evolves every single day as an arm's race and detection is a full-time job. Even a FAANG or a state actor would need to dedicate team(s) to detection technology and still have false negatives. Are people actually trying to detect AI-generated content? That would not only be pointless and futile; the threat of false positives would be enormously detrimental to anyone creating legitimate work. It is such a ridiculously bad idea I'm dumbfounded that anyone smart would be trying to do it.
- kbelder 3y agoYes, multiple teams are working on it, private and public. >It is such a ridiculously bad idea I'm dumbfounded that anyone smart would be trying to do it. Agree with you there.
- cwkoss 3y agoYep. There is a cohort of #noaiart amateur-but-wants-to-be-professional artists on twitter who believe their mediocre talents would be paying their expenses if it only wasnt for that pesky ai imggen. (Ignoring that a vanishingly small proportion of imggen art is replacing commissioned art - most is art that would simply never have been made). Like a horde of locusts, they will randomly pile onto AI artwork with hate comments for a brief period of time before moving onto the next one. People are 'offering their services' where you can DM them a link to an image and they'll eyeball it and tell you if its made by AI. Laughable hubris, if it wasn't for the inevitable ramifications of false positives.
- abejfehr 3y ago> This is just legal cover until such time as its possible to enforce no child exploitation imagery, no copyright stuff, etc. If the art used in a game violates copyright or contains imagery of exploited children, ban it of course, but what does that have to do with whether it was generated via AI or created in another manner? If anything AI generated art should be _less_ susceptible to copyrighted stuff because everything is original (even if it's not in original style)
- hot_gril 3y agoBecause AI IP law is murky, and that's all Valve cares about.