It’s not clear what would have happened if the faces were the output of a generative AI model, and didn’t actually involve real children. That seems like it's coming soon. Presumably the law is going to treat all of that the same as actual CSAM, but people will need to think hard about the reasoning for it. All of the quotes in this article stress the possibility of harm and risk to (real) minors whose faces were duplicated. Which seems like a very reasonable set of concerns!
>Presumably the law is going to treat all of that the same as actual CSAM, but people will need to think hard about the reasoning for it.
The US already signed a law in 2021 defining CSAM to include photorealistic AI generated material, even if there is no original child involved (such as deepfakes).
It feels to me that there is a (very natural!) tendency to want all of this stuff to disappear, because it’s disgusting. But that this results in a certain degree of motivated reasoning where people feel that they must justify new (invasive) legislation by tightly binding it to the welfare of real children. I worry that these justifications might ultimately do a lot of damage, causing us to aggressively treat symptoms rather than doing much to heal the underlying disease.
> Many images of child pornography have a digital fingerprint, allowing police to identify them, the judge wrote. By creating the synthetic images, Larouche made it more difficult for police to stop the spread of the illicit material.
I'm not sure where others got the idea that the underlying video wasn't child abuse material