This is a story that Hartry Field used to tell, according to Adam Elga.

A maintenance man is cleaning a laboratory when he stumbles across some brains in vats connected to computers all running a programme called "MAN DUSTS BRAINS IN VATS".

Given that the maintenance man is justified in believing that within the vats lie brains that are in the same epistemic state as he is, ought he to conclude that he is one of the brains in the vats? Field and Elga both think so, but this just doesn't make sense to me.

If he is a brain in a vat, then all of the things he is experiencing are false, or simply ideas fed into a mind by a computer programme.

If a brain and a vat are themselves just inside a computer programme then they aren't "real" brains or vats. There is nothing which is experiencing what it is like to be that brain.

He is experiencing brains in vats.

Therefore the brains in vats are false.

Therefore there is nothing that is experiencing what it is like to be one of those brains.

Therefore if he is a brain in a vat, he is not one of those brains.

Therefore this whole experience (of witnessing the brains in vats) does not increase his subjective probability that he is a brain in a vat.

I can think of two counter-examples to the above argument (or premises within it).
Firstly, it might be the case that Artificial Intelligence could be developed to the point whereby we can design computer programmes, for example simulations of the real world, that have intelligent agents within them believing it to be the real world. So often when we think of Artificial Intelligence we think of robots that have some real physical and human-like characteristics. But perhaps it would be possible to create real agents whose physical basis is a tiny proportion of a computer chip. If that were the case then the maintenance man could be a brain in a vat himself and be witnessing other brains in vats in front of him, who are experiencing life in an artificial reality within the man's artificial reality. However, if this counter-example alone were true, then his witnessing the vats would still not increase his subjective probability that he is a brain in a vat, because he still could not be those brains in vats.

Secondly, is it not possible to be a brain in a vat and still experience the real world without the need for any kind of programmable software? Suppose we take Bob's brain and stick it in a vat. The vat is attached to a computer and the computer is attached to a robotic device a few thousand miles away which has facilities for input, a camera, microphone, even cells which measure touch and pressure and heat, and the signals travel by satellite to Bob's brain where they are hooked up with his central nervous system so that he feels being pinched whenever the robot is pinched in exactly the area the robot is being pinched.

In this sort of scenario, Bob could easily be walking around the world in his robot-form and come across the laboratory with his brain in it, just like the maintenance man. If the previous paragraph is correct, that such a technology is possible, then that would refute one of our premises: the one that said that if he is a brain in a vat then what he is experiencing is not "real".

There are some people who believe that at night when they sleep, they "astrally project". What they think happens is that their consciousness leaves their body and takes on an astral body. They can then float around in the astral world. In the astral world they can see all the physical objects in the physical world, but they cannot touch them. They effectively become ghosts. They report seeing their own sleeping body, and presumably if they looked closer at their own brain, they would see the object that was experiencing everything that are currently experiencing. Now, science would say that astral projection is not a real phenomenon, that these people really only fly around a dreamworld inside their heads, not causally or physically connected to the real world at all. They wouldn't be seeing their real sleeping bodies but what they believe their real sleeping body to look like.

But I don't see how such a phenomenon could not be done with advanced technology. I've read about such an invention as a "sex suit", two people buy a sex suit and a special mannequin and put on the suit. They then have sex with the mannequin and the other person, who could be on the other side of the world, feels the impulses that the first person is giving to the mannequin, in a 2-way process. Perhaps an advanced technology would make the other person's mannequin move exactly how I am moving, so that I could effectively punch you in the face from across the world or even murder you. Some other hilarious things could occur. I could get my mannequin and throw it onto the bed, expecting to make love to my partner across the world. But unbeknownst to me, where my partner is, there is no bed there: instead there is a window. So my mannequin over there would be throwing them out of a window. Perhaps I would be a bit puzzled why my mannequin, now on my bed, is flapping its arms and perhaps even screaming, and then suddenly goes rigid and stops communicating with me.

In those circumstances, neither of our brains would be in vats. Rather, our bodies would be in a sort of vat. I have one other question to ask too. Suppose Bob was walking around New Mexico. Would it feel to him that he was really in New Mexico? My concern here is that I have a sense of self-location. I think there is a place where I really am: not my torso or my arms or legs, but in my heads, just behind my face. Now, do I have this feeling because that's where my brain is, or because that's where my eyes are? I'm not sure if Bob would feel like he's really running around New Mexico, or if he's really sitting in a vat somewhere experiencing an extremely realistic simulation of what it would be like to be running around New Mexico. I really don't know the answer to this.