I think that robots that are already fooling a few persons out of thousands of potential victims for theft are alreeady passing a weak version of Turing test, they will be better theives if they are better imitation human beings.
bad for me to says it`s way cheaper to just hire folks in india to do this. AI is a very costly bit of tech , meaning it`s doubtful it`ll ever be used to replace non-high income jobs. ex. maid service,fastfood
It is really already happening because Chatterbots don't need to be genuine AI, they are more like "expert" programs which are a large cataloge of standard responses, for a short time and within narrow circumstances theyu can pass the turing test and perhaps fool a few percent of people into opening their sensitive information. This works good and cheap and can be fired at many thousands of potential marks quickly so that a good number can be robbed before the law notices.
When genuine AI becomes availible I would expect the same cheats to use the more advanced program the same way but with faster success.
