Another year, another futurist, another ludicrous claim about robotics and articial intelligence.

Machines 'to match man by 2029' - Machines will achieve human level artificial intelligence by 2029, a leading US inventor predicts. [BBC]

A long time ago when men were men and AltaVista was the web's best search engine, this Editor was a scruffy young undergraduate at the University of Aberystwyth, specializing in artificial intelligence. The choice of degree was random, but probably strongly influenced by a love of science fiction. Unfortunately too much sci-fi can have bad consequences - some people think its real. Worse still, other people take them seriously and it ends up being top story on the BBC.

"Machines 'to match man by 2029'" was the lead story in the Science section of the BBC News website this Sunday. The whole piece was based around a couple of remarks by US inventor Ray Kurzweil, who said in a moment of optimism, "I've made the case that we will have both the hardware and the software to achieve human level artificial intelligence with the broad suppleness of human intelligence including our emotional intelligence by 2029". And that, basically, is the entire story.

On the bookshelf across the room from me sits a copy of "March of the Machines", by Kevin Warwick [1]. On the back it the blurb confidently asserts "Soon, [Warwick] will have built robots with the brain power of a cat, and in 10 years, they could be as intelligent as humans". It was published in 1997. A full review of the state of A.I. is beyond the scope of this little blog entry, but I can safely say that unless some of Kevin's Research Assistants aren't what they seem, we're nowhere near human or even cat-level intelligence yet.

Warwick and Kurzweil represents a breed of A.I. "evangelists" with a knack of getting their weird and wonderful pronouncements into the public eye. They would argue that they are popularizing science, but as Dave Green summed up in a BBC article a few years ago, "If his work is as good as he says it is, he really needs to start letting it speak for itself" [2].

There is simply no sensible basis on which to make claims about human-level artificial intelligence, when we have yet to even formally define "intelligence" or understand so much about the mind that makes us human. One day I'm sure it may be possible, but it won't be soon, and I'll gladly bet my house it won't be by 2029.

What grates though is that scientists are making these predictions off-the-cuff in order to promote a conference or sell a few more books. These are people who should be responsible, who should be presenting themselves as credible, but every time a prediction like this is made it chips away at the credibility of their fields.

That's one half of the problem - the other is that a media outlet as big as the BBC is willing to promote a non-story like this to the top of his science page on Sunday. The story isn't a report on science news, it's a single sound-bite from a guy at an AAAS meeting. At best, it's an opening line for a long and involved discussion. At worst, it's another case of the boy crying wolf.

[1]Available from 40p, but still worth more than the predictions on the cover. march of the Machines (Amazon).
[2]Why I'm Not Impressed with Professor Cyborg (BBC).