Being an early adopter was always a risk. There have always been products that did not materialize, or turned out to be crap, or that stopped working shortly after launch. It's just so much easier to reach a lot of people now, and the critical thinking abilities of entire nations is at an all time low, making for easy pickings.
I've backed a few projects, and I've been happy with them. I did not blindly back "this looks great!" projects, though. I did some research. I gauged the complexity compared to the proposed timeline and what they already had to show. I judged the amount of money they asked for up against what I thought seemed reasonable for the project goals. I evaluated whether the current state of technology seemed to even support the proposed features. There will always be risk in being an early adopter, but engaging brain before parting with money does mitigate things quite a bit.
The Humane pin was not hard to spot as high risk. Heck, anyone asking for money for anything "AI" should raise immediate red flags for everyone. I don't remember anyone being surprised when it released and was absolute crap. On the other hand, they did make an actual product of sorts. So, there's that. It's rather telling that in the world of AI products, the Humane pin is a bit of a success story in that it wasn't a complete scam.