AI-created child sexual abuse images 'threaten to overwhelm internet'
theguardian.com
theguardian.com
There’s a huge difference though. If someone listens to 90s industrial music or watches Trainspotting and shoots smack, they are harming themselves. You can’t indulge pedophilia IRL without harming someone else.
Incitement to harming others is an area where we do sometimes draw a line with speech. So I think you could argue that child porn, even if fake, is direct incitement to violence. That would put it in the same category as calls to murder Muslims or Jews, etc.
The same argument could be made about rape porn though, even if it’s fake or simulated with actors. Personally I think that anything we do conclude about CSAM should in fact apply to rape porn, revenge porn, and other porn depicting nonconsensual abuse.
So maybe we do have to tolerate fake CSAM if we tolerate an endless parade of incredibly violent war propaganda films, not to mention hate speech that goes as far as to call for persecution and violence.
My late uncle was a psychologist and always argued that our glorification of violence was far worse than almost anything in porn.
I guess maybe the problem is that if the content progression thesis holds true, some people would get more and more extreme in their pornography needs. This may be an on ramp for more child sex crimes?
a lot of brains are changing in strange ways. health issues due to consequences of climate change are rising silently and everybody knows whose fault it is. almost nothing was done for mitigation. abusers, murderers, torturers, rapists and the likes all feel the same superiority as the wealthy do making the wealthy abusers, pedophiles and so on even worse, especially in times where almost anything can be justified based on the arguments of corporate and political leadership.
But [1] from 1996 says
> There is increasing recognition that incest offenders may abuse children outside as well as inside the home. In a community sample of over 500 self-admitted male sexual offenders, Abel and colleagues(1987) found that 23.3% of men had offended against both family and non family members; 12% had offended exclusively against family members.
I'm guessing that means 65% are non-family, making your point very relevant.
1. https://www.academia.edu/download/7027369/Victims%20and%20Pe...
I did not expect machine learning to be the platform that paved the way for it. I thought we'd embed one file in an unrelated other file, and that some HMAC/AEAD algorithm would allow you to quickly know if a file has data hidden data you can decrypt. Both files being useful in their own right. Instead, we'll just flood the tubes with variations of the original, hiding the original in the crowd.
Ugh. The stuff we spend energy on producing.