In the thought experiment Mary’s Room, it is argued that Mary gains new knowledge about the colour red when she leaves the room.

It seems to me, though, that she gains no knowledge about the colour red, and instead only knowledge about herself (ie. how does Mary’s brain respond when it perceives the colour red).

And, of course, such knowledge could be obtained without access to the colour at all, simply by attaching electrodes to the nerves in the eye to stimulate the nerve cells would be sufficient.

Finally, consider an artificial intelligence in the same room with the same rules. Except that, unlike a human, this intelligence can simulate precisely the input to its visual system based on its understanding of phenomena. In the thought experiment, the understanding would be complete, and thus the simulation would be perfect. So, the artificial intelligence would have no new stimulation when leaving the room.

The only substantive thing the thought experiment seems to actually argue is that humans are incapable of achieving “understanding” through purely knowledge-based transfer. Though it does nothing to argue that that limitation is a universal constant instead of an accident of the evolution of the human brain.