A computer has passed the Turing test for humanity – should we be worried?

http://www.theguardian.com/commentisfree/belief/2014/jun/13/computer-turing-test-humanity

Version 0 of 1.

Apparently, a super-smart computer has just passed some geeky game of "knock, knock, who's there?" designed by Alan Turing back in the 1950s. "A computer would deserve to be called intelligent if it could deceive a human into believing that it was human," said Turing. And now some say it has done precisely that, recently fooling a number of judges by successfully impersonating a 13-year-old boy. Not that this thing popped round to drink Coke, watch the football and swear at the referee. The test is nothing more than a five-minute exchange of texts.

This is light years away from a computer "being human". To do that, claimed sci-fi author Philip K Dick, it would need to look like us, walk and talk and act like us, and successfully mimic our emotional and empathetic responses. In Bladerunner, that cult 1982 classic based on a Dick short story, Harrison Ford's job is to track down and "retire" rogue replicants. The existential problems begin when he falls in love with one of them. The question posed is not so much whether the beautiful Rachael is nothing more than a squishy computer with red lipstick, but whether we are anything more than that ourselves.

Like all sci-fi, Bladerunner tells us more about the time that it was made than the future it imagines. Set in LA in 2019, now only five years hence, the world of replicant humans no longer looks like it is before us – not even remotely so. According to Moore's law, computers are supposed to get twice as smart every two years, yet they have not become nearly as smart as us dumb human beings. The point being, of course, that we do different things – are different things.

I am suspicious of big questions like "what is it to be human?", especially when they seem to have slipped their moorings from the context of ordinary speech. For when we ask "can't you see this is a human being?", we are not under the impression that there is any confusion involved. If we ask this of a torturer, we don't suppose he has made a philosophical mistake. Rather we are making a demand for his victim to be treated with a certain sort of moral respect. Likewise, if we cruelly say that someone is "not human", we are usually claiming that they do not deserve such consideration. In other words, to claim something as human is to make a moral demand that they are handled with care. Thus, for instance, when Turing was chemically castrated as a cure for his illegal homosexual activity, we now rightly regard his treatment as inhumane. But in saying this we are not saying that the state made some sort of category error. Rather, that the state was cruel, that its blindness was moral. In other words, being human is fundamentally a moral category not a biometric one. This is the better Turing test for humanity.

So why, you may ask, do human beings deserve this moral respect? It's not because they can think. Or because they have language. Or even that they have souls, whatever one of those is. Indeed, it's not because of anything about them other than that they are human. In other words, the essence of the category human is that it is something that demands moral respect. And why? Answer: simply because humans are human. This circularity feels wholly unsatisfying because we labour under the impression that a moral attitude requires grounds beyond itself. But understood thus, every answer to a why question gives rise to another deeper why question, thus to an infinite regress. Where then do the questions come to an end? Generally, in anything other that science fiction, being human is where the questions stop. This is as deep as it gets.

Twitter:@giles_fraser