As a UK-based amateur photographer(for the past 51 years) I'm extremely disappointed that AI is turning out to be a threat to my exclusive right of ownership of the photos I've taken and, since the advent of the digital age, posted on the internet.
I've paid for my camera equipment, my travel, accommodation, photo processing and post-processing software. I own the copyright to my photos and, therefore, I believe that I should retain exclusive rights to who sees them and how they are used. If anyone wants to use them, they should, in my opinion, have to request my permission formally and in writing. I would also expect some form of 'quid pro quo' and guaranteed limitation as to usage in return.
I get that AI needs to 'learn' from somewhere. However, that 'somewhere' should surely be stock image libraries where the image owners will be compensated for the AI algorithm scraping them from the internet - and not those of us amateurs who have spent time and not a little money honing our craft and producing things which are, and should remain, exclusively ours.
Adobe tried to pull this stunt in 2024 and was told to sod off. They had to back down. The UK government's intention might embolden them to try again, though I hope not.
There is a little hope, though. There are some software programs which seen to protect our images from just this sort of threat but acting as a 'cloaking' device against the AI bots searching out and copying our work to assimilate via machine learning. Here is one example:
I haven't used it, yet, and I don't know how effective it is, but I hope it works and I hope that it becomes the tip of a huge iceberg of such software that can be used to scupper the theft of our work, at least in some significant way - if not totally.
I really don't have answers to those questions, Graham; nor do I know what the Charting community is. I saw reference to the UK government's intentions on AI in the press recently and wondered if any Substack articles had mentioned it. That's when I came across this one and decided to say my two-penneth. I'm not anti-AI when it comes to what it can do FOR us but I'm 100% against it when it's being used (or being allowed to be used) AGAINST us. Double-edged sword, I know. The Glaze website suggests it is a useful tool. If the likes of Adobe, Capture One and the other post-processing software manufacturers would see their customers as being the people who provide the funding for their continued existence, we might be able to push back. However, I'm going to start 'glazing' the photos I produce before posting them and do my tiny bit to spread the word and disrupt this move.
Hi Paul, that was a question to other readers in the Charting Gen AI community (I abbreviate that to Charting!). Johan Brandstedt has just posted this reply to a Substack note on this: "Using both that and Nightshade. Glaze has been broken by an adversarial researcher. This doesn’t mean it’s useless, as there is still no surefire way of detecting it afaik, and as it still imposes removal cost and removes plausible deniability on attackers."
Thanks for that, Graham, and thanks also to Johan Brandstedt. I've just checked out 'Nightshade'. It appears to be from the same stable as Glaze but, whereas Glaze is a defensive tool used individually by artists / photographers, Nightshade is "an offensive tool that artists can use as a group to disrupt models that scrape their images without consent (thus protecting all artists against these models)." Maybe this is something that social media sites might consider using across whole platforms to fend off attacks from AI "researchers"?
As a UK-based amateur photographer(for the past 51 years) I'm extremely disappointed that AI is turning out to be a threat to my exclusive right of ownership of the photos I've taken and, since the advent of the digital age, posted on the internet.
I've paid for my camera equipment, my travel, accommodation, photo processing and post-processing software. I own the copyright to my photos and, therefore, I believe that I should retain exclusive rights to who sees them and how they are used. If anyone wants to use them, they should, in my opinion, have to request my permission formally and in writing. I would also expect some form of 'quid pro quo' and guaranteed limitation as to usage in return.
I get that AI needs to 'learn' from somewhere. However, that 'somewhere' should surely be stock image libraries where the image owners will be compensated for the AI algorithm scraping them from the internet - and not those of us amateurs who have spent time and not a little money honing our craft and producing things which are, and should remain, exclusively ours.
Adobe tried to pull this stunt in 2024 and was told to sod off. They had to back down. The UK government's intention might embolden them to try again, though I hope not.
There is a little hope, though. There are some software programs which seen to protect our images from just this sort of threat but acting as a 'cloaking' device against the AI bots searching out and copying our work to assimilate via machine learning. Here is one example:
https://glaze.cs.uchicago.edu/index.html
I haven't used it, yet, and I don't know how effective it is, but I hope it works and I hope that it becomes the tip of a huge iceberg of such software that can be used to scupper the theft of our work, at least in some significant way - if not totally.
Thanks for the comment Paul. Wondering if anyone else in the Charting community has used Glaze? How effective is it?
I really don't have answers to those questions, Graham; nor do I know what the Charting community is. I saw reference to the UK government's intentions on AI in the press recently and wondered if any Substack articles had mentioned it. That's when I came across this one and decided to say my two-penneth. I'm not anti-AI when it comes to what it can do FOR us but I'm 100% against it when it's being used (or being allowed to be used) AGAINST us. Double-edged sword, I know. The Glaze website suggests it is a useful tool. If the likes of Adobe, Capture One and the other post-processing software manufacturers would see their customers as being the people who provide the funding for their continued existence, we might be able to push back. However, I'm going to start 'glazing' the photos I produce before posting them and do my tiny bit to spread the word and disrupt this move.
Hi Paul, that was a question to other readers in the Charting Gen AI community (I abbreviate that to Charting!). Johan Brandstedt has just posted this reply to a Substack note on this: "Using both that and Nightshade. Glaze has been broken by an adversarial researcher. This doesn’t mean it’s useless, as there is still no surefire way of detecting it afaik, and as it still imposes removal cost and removes plausible deniability on attackers."
Thanks for that, Graham, and thanks also to Johan Brandstedt. I've just checked out 'Nightshade'. It appears to be from the same stable as Glaze but, whereas Glaze is a defensive tool used individually by artists / photographers, Nightshade is "an offensive tool that artists can use as a group to disrupt models that scrape their images without consent (thus protecting all artists against these models)." Maybe this is something that social media sites might consider using across whole platforms to fend off attacks from AI "researchers"?