Friday, 1 April 2016

Machines "not something to be feared"?

On Friday, 25 March 2016, The Straits Times ran a story with the headline "Machines 'not something to be feared'". The basis of this headline was an interview with Mr Demis Hassabis, co-founder and chief executive of DeepMind, the company that created AlphaGo, which the previous week beat world Go champion Lee Se Dol 4-1 (correction from earlier post). I disagree with Mr Hassabis' prognosis.

Mr Hassabis says: "In the next five years, it would be great to see machine-learning [the distinctive achievement of AlphaGo] applied to healthcare in a deep way for medical diagnosis." Well, that's one scenario.

What about the dystopian scenario that machines are "something to be feared" and will "take over the world and wipe out humanity"? Here is Mr Hassabis' response: "There are these science fiction scenarios but they're just science fiction. I don't think we should confuse Hollywood and what's really reality."

Science fiction has a habit of coming true. Witness telephones, television, satellites, and yes, even computers. Relegating the dystopian scenario to science fiction actually supports the dystopian scenario rather than rebuts it.

The entirely material AlphaGo computer demonstrated all the intuitiveness, creativity and innovation -- all up till now declared as uniquely mental attributes -- needed to defeat a human mind. The machine has passed the Turing Test. The machine is intelligent, never mind the prefix "artificial".

At some level of complexity, the machine may well declare as Descartes did some centuries ago: "Cogito ergo sum", "I think, therefore, I am". The machine will declare its intelligent consciousness, and hence its existence.

It is a short logical step from that to declaring that it is alive, that it is a life, that it has rights. After all, the argument is widely accepted that animals have rights just because they are alive and feel pain, the much stronger argument will be that an intelligent and conscious "machine" (a word soon to be a misnomer) also is alive and has rights.

What philosophical recourse will there then be to distinguish human life from machine life? To appeal to physical form would not just be facetious; indeed, it is easily overcome -- with improved robot design. Japan, for example, is replete with humanoid robots performing previously human functions.

We already have CAD, computer-aided design. With further development of machine-learning, the day will come when we will witness computer design and manufacture. Let me put this bluntly: Machines will learn to reproduce themselves. A new species will well and truly arrive on Planet Earth.

Cue Darwin and evolution. Call this dystopia? No, it is the future.

Cheers.

No comments: