General Discussion

There's 1000's of companies using custom models built to do business-specific tasks that are in NO WAY trained on stolen data. When you talk to the chatbot at the cable company or whatnot, that is AI, but it's not any of the big AI companies models that used stolen data. That AI just needs to know enough to do it's job (which admittedly they often do fairly poorly). The shop where I work put AI into a project we did for a client that was trained on NOTHING but the companies own data/information, in order to detect fraudulent referrals (this was for a company that pays out for "affiliate links" ).

That's what you're missing in this equation