Prompt's TechBlog

Last night's Horizon programme on BBC 2 was all about artificial intelligence, a subject close to my heart, so I was quite interested to find out whether the programme makers would do the topic justice, or if we'd be treated to an hour of insensible pseudo-science from the likes of Kevin Warwick. Fortunately Captain Cyborg didn't make an appearance in the show, but while some genuinely interesting areas were touched upon, these were interspersed with the kind of ridiculous apocalyptic scaremongering which now seems to go hand in hand with any media coverage of AI.

The general theme of the show was that computing power and AI research is accelerating at such a pace that it now seems plausible, even likely, that within our lifetime we will reach the Singularity, the point at which computers will equal human intelligence. If you can match human intelligence, you can also exceed it, so it stands to reason that once we reach that point, we will be able to build machines that are far, far more intelligent than us. Godlike, as the show's narrator repeatedly pointed out.

The show featured interviews with plenty of reputable people working in this field, including Ray Kurzweil, and staff at the Brain Mind Institute, so it was shaping up to be an interesting and rational examination of the topic. But then it all went a bit wrong, the narrator started telling us that creating highly intelligent computer systems would result in people being turned into super-human cyborgs. And, of course, there was the unquestioned assumption that any such highly advanced computer system would immediately go mental and make wiping out humanity its first priority. This seems like something of a confused leap to me:

Scientist A: "Congratulations professor, we've finally created a computer with super-human intelligence. This is a truly historic moment! Which of the great scientific challenges shall we set it to work on?"

Scientist B: "I've got a better idea, let's all get cybernetic brain implants instead, so it can control our minds and ultimately destroy the human race."

Scientist A: "Yeah, ok, why not. What's the hotkey to start the 'Crush Humanity' sub-routine on this thing?"

Then they trotted out Hugo de Garis, an AI researcher with something of a 'science fiction' attitude to it all. According to Professor de Garis mankind faces a brutal and inevitable all out war in which supporters and opponents of AI (cosmists and terrans, as he calls them) will fight to the death for control of civilisation. I'm not making this up.

Next up the program makers threw in a thread about the Unabomber, pointing to this lone wacko's bombing campaign against universities and angry anti-technology manifesto as proof that the AI wars had already begun. What should have been an interesting investigation into the current state of AI research turned into the usual sci-fi rubbish, somehow equating the development of smart computers to the end of the human race. The point they never seem to make in these programs is that if the computers start getting too uppity, we can always just flick the off switch, which is what I should have done to my TV last night.