Well, the thing about fraud is that it’s detectable by intelligence. So from your original post, it could be that middle-intelligence people can use AI to defraud others. But, IMO it won’t largely be successful on highly intelligent people, especially those who themselves adopt AI, if that makes sense.

There’s nothing stopping people from using AI to detect and combat fraud, just as people adopt AI to perpetuate fraud. In fact I would guess that the power dynamic would be in the defenders’ favor. It’s hard to convince an AI to help you commit fraud. It’s easy to convince them to help you detect and mitigate it.