"Over 90% of AI-generated images are sexually explicit" This is pretty much bullshit. This is easily debunkable just by looking at the sheer amount of content on the Internet which is *not*.
The "source" is a "report" from *2019* (from a British org) which doesn't look into AI generated content, it looks at content branded as being "deepfakes".
The same applies to the "non-consensual" point. If someone is investigating the use of some "deepfake" tool, that is different from an "AI generation tool".
The "source" is a "report" from *2019* (from a British org) which doesn't look into AI generated content, it looks at content branded as being "deepfakes".