I just won a chargeback dispute initiated by Muse.
A person bought a thing from me, then bought the same thing 48 seconds later. I can tell from the request log that the person hit the confirm button multiple times.
No big deal. I have a little script that monitors for things like this. Or they could have emailed for a refund. Very easy to resolve.
Instead, their Muse bot initiated a chargeback, almost certainly after they approved it.
The chargeback costs me money, so I asked my customer to cancel it.
They responded with an apology and described that Muse did it, not them!
So, I sent that explanation to the bank, which costs another $50, and I won, and got all of my money back.
Folks are (rightly) worried about AI (security vulnerabilities, deep fakes, scams, etc.), but another way Skynet takes over is by creating a mountain of shit that causes work in the real world.
AI slop requiring manual review in the court system is one example.
My chargeback is another.