In March, a girl’s stepfather took his own life after cops discovered that he had used Grok to create 7,000 sexually explicit images using one photo taken when his stepdaughter was 11 years old, the amended complaint alleged.
Grok allowed the man to generate extreme images depicting incest and rape without flagging any harmful behavior, the complaint said. Seemingly, xAI’s child safety system only intervened after the man input a prompt for “gang rape.” That request sent a CyberTip to the National Center for Missing and Exploited Children (NCMEC), which alerted law enforcement to the AI CSAM.
Yet the harm was not stopped then, either. Despite mandatory reporting requirements to share information like a user’s IP address when CSAM is flagged, xAI repeatedly refused to help cops or NCMEC identify the user, the complaint alleged. For weeks, xAI allegedly “obstructed this investigation at every turn” and made it harder for “law enforcement efforts to locate, identify, and apprehend the perpetrator.”
Eventually, the stepfather was arrested after cops obtained a warrant to seize his devices. That’s when “a forensic review revealed approximately 7,000 AI-generated images and videos” depicting his stepdaughter, which were allegedly produced using Grok. Without Grok providing users with easy access to “undressing” capabilities, his family doubts he ever would have generated the harmful images, which he allegedly trafficked online in trade for “CSAM produced by other child sex predators.”
I think we'll never know scientifically because there's no ethical manner of conducting the study. No one's going to supply a known pedophile with CSAM and see whether or they they molest someone, or hook up a gizmo to their head and measure their brainwaves when they go to a playground. It would probably answer a lot of questions if they did, but we as a society usually agree these questions aren't worth scarring a child over.
I’d argue for using a study with something like another common, but legal and harmless paraphilia - feet was the one I suggested earlier - but otherwise arbitrary. I’m making an argument based on transferability but that’s reasonable here I think. The only thing we can’t control is societal disapproval, so maybe you can substitute something else like hardcore BDSM, but you can’t really replicate the pure evil factor of CSAM. Maybe animal abuse content in an area where that is legal but I don’t really like that either. Rough misogynistic BDSM content is probably the happy medium here.
You’d do something like get two groups. One that you prevent from accessing the content somehow, and one that has access to that content. The obvious problem is controlling access to content. Maybe it’s just a monitoring program with unrestricted access to the internet but tracking visit frequency to that type of content, where one group is told/incentivized to not visit such content, while the other is given no such instructions; maybe it has to be self reported frequencies. Maybe one group is explicitly given just the instructions that specific type of content is specifically legal and that there is no harm in enjoying that content. Recruit among communities that enjoy that specific content, in forums where they are already not self censoring.
Maybe another DV is ranking subjective scores of intensity. I think you could do this by categorizing tags, which are amply available on most pornographic content, into sorta but arguable subjective ratings of intensity. Maybe this is a mixed methods paper where the quantitative aspect is the frequency of visits stuff, while qualitative is theme analysis on the tags.
If it has to be specific to CSAM, perhaps it’s measurements of how many possessors go onto be offenders over their life time. I doubt mandated counseling works often enough to be that much of a confounding variable.
Well, that's kinda a problem, because what if it's one way for a harmless paraphillia, but different for a harmful one? You'd never know without doing that crucial and specific experiment. You could at best extrapolate that it could plausibly be the same, but that's not the same as actually knowing.
I apologize if it looks like I'm being pedantic. I'm in kind of a philosophical mood. I felt the need to clarify I'm not trying to be difficult for whatever reason.
No research can prove a positive.
Most research in social sciences uses a p-value of 0.05. Statistically, this means that we are expected a 1/20 chance of a type 1 error, a false positive.
It’s more about establishing patterns than “proof” of any kind of absolute law. Human behavior isn’t governed like rigid bodies or genetics.
I'm more into art and philosophy than science, so if that's your background then I'm not going to argue with you. It's obvious from the way you talk versus how I talk that we're approaching the question from fundamentally different angles. I'll kindly return to up my own ass, where all philosophers go when we're done talking.