Re-enter The Singularity – and its déjà vu all over again

I see that Google and NASA are financing a school for futurists in Silicon Valley ‘to prepare scientists for an era when machines become cleverer than people’, as the Financial Times puts it. This ‘Singularity University’ will teach its students about nanotechnology, AI and that sort of thing. It also involves Ray Kurzweil, the inventor, enthusiast for AI and inventor of the term ‘singularity’, which he uses to describe the moment when the world will change because machines are suddenly smarter than people. At that point, the machines will solve all our problems, including climate chaos, peak oil and world hunger. And no doubt cancer and the common cold too.

According to the LA Times (on February 2 2009):

The goal of the Singularity school, which will be located at an Ames facility in Sunnyvale, Calif., is to bring together the world’s top graduate and postgraduate students in 10 diverse disciplines, such as biotechnology, nanotechnology, medicine and law. They will spend nine weeks together learning about each others’ disciplines and then focus on… finding ways to overcome pressing challenges such as poverty, hunger and pandemics.

Let’s get some basic facts straight.

Firstly, there is not one machine in the world that comes close to passing the Turing test for artificial intelligence. Now bearing in mind that the Turing test is singularly undemanding – basically the ability to fake intelligence for five minutes – you would have thought that if ‘the singularity is near’ (one of Kurzweil’s book titles) it would be pathetically easy to pass. After all, it was invented in the 1940’s, when the sum total of the world’s computing power was less than what is sitting on my desk right now. Yet so subtle is intelligence that no one has yet come close to claiming the prize.

Secondly, although it is the case that the total processing power of all the computers on the planet is truly vast, all the computers on the planet have exactly as much intelligence as a gatepost. They may be fantastically useful for all the purposes noted above, but they will perform this invaluable service without showing a spark of intelligence. The emergent properties effect of technologies such as the internet and nanotechnology may be very surprising indeed, but there seems to be no specific reason to believe that any of them will constitute intelligence proper, and more than piling up rocks, though it may result in something as wonderful as Chartres cathedral, will ever result in an organism.

In summary, I think a more archaic response to this Singularity University is called for here, and that is to say, quite simply, ‘bollocks’.

As I have argued in detail elsewhere in this blog, it is intrinsically impossible for any account of intelligence that is inspired by AI or ‘cognitive science’ (was ever a discipline worse named?) to analyse the basic facts of real intelligence. Concepts such as computation and information processing are logically incapable of grasping what it means to be intelligent. So constantly harping on about artificial intelligence in this particular vein is a fool’s errand.

Of course, we’ve been here before – apparently inspired/possibly mad inventor proclaims the dawning of the Age of machine Intelligence, and in the absence of a decent analysis of what intelligence actually means, we all bow down. But then it turns out to be just another remake of a much older story – the Emperor’s New Clothes.

More of RJ Robinson at http://richardjrobinson.blogspot.com/

Leave a Reply