In his Reddit AMA, David Eagleman, neuroscientist/writer/host of PBS show The Brain, was unsurprisingly asked one of the key questions in his field, whether machine intelligence can ultimately become conscious. The excerpt:
Question:
Does it seem plausible to you that consciousness could arise from a machine/computer? Is this not all the brain itself is?
David Eagleman:
It does seem plausible to me that consciousness could arise from a machine. However, we should note that this is just a hypothesis, and none of us yet know if it’s true. The idea is called the Computational Hypothesis, which proposes that the mind emerges from the interactions of the billions of neurons over their trillions of connections. Implicit in this idea is that the mind is the software that runs on top of the hardware of the brain, and that the same software could be run on any substrate, from beer cans and tennis balls to zeros and ones in a computer. In this view, there’s nothing special about the biological wetware — all that matter are the computations that run on top.
I would say that this is the assumption many of us work under, but not all neuroscientists agree. Some very important thinkers (e.g. Christof Koch, David Chalmers) argue that consciousness will not emerge from a computer simulation, because consciousness may be a property of matter that cannot be simulated, in the same way that a weather simulation does not actually get wet (see panpsychism).•