@Dolimac@ArtExplosionNFT@ironspider_952@loishh Don't know, didn't ask. Proof of inquiry and purchase is enough, at least in my home country. Copyright laws vary from country to country.
@PatriotsOne87 @dhaliwal1724 @theJagmeetSingh Here it costs me $75 if I want to upgrade to the 4gb data plan with unlimited texts and calls. Most of the time I don't get coverage in certain towns or areas that aren't even in the middle of nowhere.
@PatriotsOne87 @dhaliwal1724 @theJagmeetSingh Canada has the highest prices for absolutely no reason. Our 5G isn't better than Germany's or whatever other country. When I went to Italy I got an unlimited calls/texts/data for $18 with Vodafone including roaming with 27 European countries.
@epicRedHot1 @artsycookies_ No, you need to open the file and anise the metadata. If the metadata doesn't I donate anything relevant (in case of a screenshot), then the engine will think the image is public domain and feed it to the AI. Unless another alternative is proved. We are talking pixels.
@epicRedHot1 @artsycookies_ This is an AI generated image (random prompt btw), whatever model comes next can and will trained from this image which was trained based on something else in a "Russian doll" fashion.
@epicRedHot1 @artsycookies_ Not only this but training softwareres are not only pulling from all sorts of images (copyrighted or not) but they're also training by pulling images from themselves which are all generated and marked as public domain (according to metadata). These images are out there right now.
@RagingOwlbear@AndySemetic14 Regardless, models aren't designed nor trained to retraced the steps used to generated images. Meaning there's not guaranteed it'll find wether the image is based on copyrighted material or not.
@RagingOwlbear@AndySemetic14 Right but to "test" it requires GPU power which is very expensive, for each test it costs as much as 25c USD (the cost of a single generation according to OpenAI). Considering all the amounts of prompts requested per hour, that's a hefty amount of power.
@RagingOwlbear@AndySemetic14 What tool do we have at our disposal to determine which image has been created by taking information from non consensual spruces? Like this one for example. Can we, in detail, determine weather or no it is generated based on copyrighted material?
@RagingOwlbear@AndySemetic14 I think AI models must be trained using free sourced data. 100% agree with you on this. I do have a question. What is there to do with open sourced models that deliberately train AI to use copyrighted materials and are freely distributed?
@RagingOwlbear@AndySemetic14 One of the most educating and rational repsonses I've received today. I did not know that, and I thank you for the information.