I wonder, with how aggressive most copyright holding companies are, if there is going to be a push to hold AI companies liable for reverse engineering like this. It possible AI companies designate it as forbidden knowledge, and “realign” models to not allow it