I could give you some great horror stories, but honestly I don't see the benefit in either potentially harming former coworkers of mine that still work at those places or ending myself in some sort of career/legal trouble for something people generally don't care about (other than a few points on HN).

If you were caught demoing something both horrific and internal you would risk serious damage to your career, and ultimately will have zero impact on the industry as there's just too much data out there and too much money wrapped up in it.

Plus, most people working with the data don't bother to look at it. The places I've internally demo'd massive privacy risks were shocked because they didn't realize what their own data was capable of. Most people are just writing jobs that run and shuffle data around from one place to another never really asking "what is this data?" Even among data scientists I'm routinely surprised (so maybe I shouldn't be surprised) how frequently data scientist never do any real error analysis by looking at what the model got wrong and trying to understand why.