An 'antidote' to the recently released AI poison pill project known as Nightshade. - GitHub - RichardAragon/NightshadeAntidote: An 'antidote' to the recently released AI poison pill...
I don’t think most people are collecting images by hand and saying “ah yes I’m just gonna yoink this and use it in my model”. There are a plethora of sites for sharing repositories of training data, and therefore it’s pretty easy for someone training a model to unknowingly pull down some data they don’t actually have permission to use. It’s completely infeasible to check licensing by hand on what could be millions of images, so this tool makes it easy to simply not train on images that have gone through Nightshade. I fail to see how that’s unethical, as not training on the image is the whole reason the original image was put through Nightshade in the first place.
Then it shouldn’t be done. That’s the unethical part. Trying to just avoid the problem by continuing to scrape large data sets for images that you shouldn’t be using is the entire problem. Either get permission for each image or don’t build your image model. Doing otherwise is unethical.
Again, in many instances, folks training models are using repositories of images that have been publicly shared. In many cases the person/people who assembled the image repositories are not the same person using them. I agree that reckless scraping is not responsible, but if you’re using a repository of images that’s presented as ok to use for AI training, I’d argue it’s even more ethical to strip out the Nightshaded images, because clearly the presence of Nigthshade means you shouldn’t use that one. I guess we’re just going to have to agree to disagree here, because I see this as a helpful tool to specifically avoid training on images you shouldn’t be.
I don’t think most people are collecting images by hand and saying “ah yes I’m just gonna yoink this and use it in my model”. There are a plethora of sites for sharing repositories of training data, and therefore it’s pretty easy for someone training a model to unknowingly pull down some data they don’t actually have permission to use. It’s completely infeasible to check licensing by hand on what could be millions of images, so this tool makes it easy to simply not train on images that have gone through Nightshade. I fail to see how that’s unethical, as not training on the image is the whole reason the original image was put through Nightshade in the first place.
Then it shouldn’t be done. That’s the unethical part. Trying to just avoid the problem by continuing to scrape large data sets for images that you shouldn’t be using is the entire problem. Either get permission for each image or don’t build your image model. Doing otherwise is unethical.
Again, in many instances, folks training models are using repositories of images that have been publicly shared. In many cases the person/people who assembled the image repositories are not the same person using them. I agree that reckless scraping is not responsible, but if you’re using a repository of images that’s presented as ok to use for AI training, I’d argue it’s even more ethical to strip out the Nightshaded images, because clearly the presence of Nigthshade means you shouldn’t use that one. I guess we’re just going to have to agree to disagree here, because I see this as a helpful tool to specifically avoid training on images you shouldn’t be.