I’ll admit I tried talking to a local deepseek about a minor mental health issue one night when I just didn’t want to wake up/bother my friends. Broke the AI within about 6 prompts where no matter what I said it would repeat the same answer word-for-word about going for walks and eating better. Honestly, breaking the AI and laughing at it did more for my mental health than anything anyone could have said, but I’m an AI hater. I wouldn’t recommend anyone in real need use AI for mental health advice.

Fakespot was somewhat accurate at catching when Amazon sellers take a well-reviewed item and swap out the product for another, by changing the title, description, and pictures. We’ve probably all read a review on Amazon that feels like the reviewer is posting a review of a completely different product, like a review that seems to be about a kitchen utinsil on a listing for an unusually affordable camera. It’s a pretty common scam that Fakespot was pretty good at catching. It didn’t seem as good at adjusting ratings for legit products and seemed to kind of randomly knock off a a half to one and a half stars on pretty much every listing, even on quality products.