Yes. You're saying that the AI trainers must have had CSAM in their training data in order to produce an AI that is able to generate CSAM. That's simply not the case.
You also implied earlier on that these AIs "act or respond on their own", which is also not true. They only generate images when prompted to by a user.
The fact that an AI is able to generate inappropriate material just means it's a versatile tool.
This comment thread started with you implying that the AI was trained on illegal material, I'm really not sure how it's got to this point from that one.
AI only knows what has gone through it's training data, both from the developers and the end users.
Yes, and as I've said repeatedly, it's able to synthesize novel images from the things it has learned.
If you train an AI with pictures of green cars and pictures of red apples, it'll be able to figure out how to generate images of red cars and green apples for you.
I don't know what specifically Microsoft is planning here, but in the past I've taken screenshots of my settings window and uploaded it to Copilot to ask it for help sorting out a problem. It was very useful for Copilot to be able to "see" what my settings were. Since the article describes a series of screenshots being taken over time it could perhaps be meant to provide context to an AI so that it knows what's been going on.
You realize that there are perfectly legal photographs of female genitals out there? I've heard it's actually a rather popular photography subject on the Internet.
Do you see where I'm going with this? AI only knows what people allow it to learn...
Yes, but the point here is that the AI doesn't need to learn from any actually illegal images. You can train it on perfectly legal images of adults in pornographic situations, and also perfectly legal images of children in non-pornographic situations, and then when you ask it to generate child porn it has all the concepts it needs to generate novel images of child porn for you. The fact that it's capable of that does not in any way imply that the trainers fed it child porn in the training set, or had any intention of it being used in that specific way.
As others have analogized in this thread, if you murder someone with a hammer that doesn't make the people who manufactured the hammer guilty of anything. Hammers are perfectly legal. It's how you used it that is illegal.
You suggested a situation where "many people would get off charges of real CSAM because the prosecuter can't prove that it wasn't AI generated." That implies that in that situation AI-generated CSAM is legal. If it's not legal then what does it matter if it's AI-generated or not?
First, you need to figure out exactly what it is that the "blame" is for.
If the problem is the abuse of children, well, none of that actually happened in this case so there's no blame to begin with.
If the problem is possession of CSAM, then that's on the guy who generated them since they didn't exist at any point before then. The trainers wouldn't have needed to have any of that in the training set so if you want to blame them you're going to need to do a completely separate investigation into that, the ability of the AI to generate images like that doesn't prove anything.
If the problem is the creation of CSAM, then again, it's the guy who generated them.
If it's the provision of general-purpose art tools that were later used to create CSAM, then sure, the AI trainers are in trouble. As are the camera makers and the pencil makers, as I mentioned sarcastically in my first comment.
The person who was charged was using Stable Diffusion to generate the images on their own computer, entirely with their own resources. So it's akin to a company that sells 3D printers selling a printer to someone, who then uses it to build a gun.
No, you keep repeating this but it remains untrue no matter how many times you say it. An image generator is able to create novel images that are not directly taken from its training data. That's the whole point of image AIs.
The trainers didn't train the image generator on images of Mr. Bean hugging Pennywise, and yet it's able to generate images of Mr. Bean hugging Pennywise. Yet you insist that it can't generate inappropriate images without having been specifically trained on inappropriate images? Why is that suddenly different?
The state of the art for small models is improving quite dramatically quite quickly. Microsoft just released the phi-3 model family under the MIT license, I haven't played with them myself yet but the comments are very positive.