@caludio @saagar Are we trusting it to do so?

Right now -- and probably forever -- AI's big problem is reliability. It can give you *an* answer, and under some circumstances a *right* answer, and security is so hard for humans because that's mostly not good enough.

AI seems to have a very human-like security problem profile. Which is a bad sign in this specific case.

@codefolio @caludio AI is IMO worse than humans currently because it is quite easy to trick it, far easier than it is to trick the average person

@saagar @caludio @codefolio easier to trick but on the other hand humans have limited endurance

@roboneko @caludio @codefolio Yes and in that aspect AI is much better. However replacing your fraud prevention person with a hundred children is unlikely to actually be a positive step even though they might be able to review more cases

@saagar @caludio @codefolio ok but what if we have the children flag things they think are really suspicious and otherwise leave the adult alone

I'm not saying AI is there just yet but it seems (to me at least) like a plausible next-year-or-so development

@roboneko @caludio @codefolio idk my experience with children is that they never leave adults alone except when they are doing something contrary to what the adults told them to do