Alex Garland’s film Ex Machina seems to have prompted a lot of discussion about the progress being made with artificial intelligence, and whether it will ever supersede human life on earth.
Of a number of articles in the press, Nicholas Carr’s in The Guardian was perhaps the most thoughtful. The debate falls into two camps: those who believe it is only a matter of time before A.I. becomes capable of outsmarting humans, and those who believe it never will.
Carr identifies the problematic non-transferability of artificial intelligence – in other words, A.I. can be vastly smart at what it is programmed to do and simultaneously hopeless at anything outside that realm. In the wrong hands, such single-mindedness could be lethal.
He also proposes that what makes humans smart is not their ability to process vast amounts of hard data but their ability to make sense of things, drawing on not only information but also observation, prior experience and emotion, and then weaving them into a whole, in a way that permits us to respond to the world in a manner both more sophisticated and subtle – and less predictable – than any machine. It’s going to take a formidable machine to equal the 100 billion neurons in the human brain.
This is why I think Carr is right when he says that the advantage we have over machines is that we are alive and they are not. The important thing is the fluidity of thought that those 100 billion neurons permit. Machines may become better than humans at specific tasks – but as Rhodri Marsden observed in The Independent, while we might end up with a cyborg that can paint like Monet, the chances of its also being able to come up with Duchamp’s Urinal are pretty remote.
In a sense, we are once again discussing the concept of causal density – the idea that reality is so complex as to be unpredictable. And the human mind is part of that. What makes us human is not our ability to be rational, but to go beyond that, into the realms of creativity, imagination, empathy and emotion. Machines can ape some human emotions, but that’s about as far as it goes – and as far as we know, they don’t actually experience them.
Carr suggests that a greater threat is becoming too dependent on A.I., such that we eventually lose those higher abilities. There is some evidence of that already, and I think there is more emerging in schools, where we might expect to see the first impacts of new technologies on up-coming generations. In particular, I am thinking about the decline I perceive in manual dexterity, including handwriting and general graphicacy.
Carr also discusses MOOCS (Massive Open Online Courses), which he claims are not being as successful as was predicted. He argues that this is due to their failure to replicate one aspect of traditional teaching: the largely indefinable effect of putting real human beings together and getting them to interact; here is another way in which the human sum becomes greater than its parts, and it’s why I believe that human teachers will always be, if not technically needed, at least desirable in human terms. Even if we can produce machines that one day can replicate such traits, I have my doubts that they will interact with humans fully successfully, simply because people will never trust a machine in the way they trust another human being.
Carr ends by suggesting that we should respect the abilities of smart machines, but that we should respect human capabilities even more. It was at this point that I saw a further irony: at the same time as we are up-skilling machines, we seem to be deskilling humans. Education is increasingly being seen merely as an exercise in logic and technical proficiency. The running is being made by the scientists and mathematicians within the education sector, whose concern is (rightly) with the transmission of technical skills, but whose model is being projected onto education as a whole. Yet in conversations, I am often left sensing that such people sometimes have even less real appreciation of, or time for, the more subjective – I would say humane – aspects of life than my humanities-derived prejudice suggests.
And yet it is these unpredictable and often creative aspects that form the core of what it is to be human. The majority of people’s lives are, I would suggest, lived more as an emotional narrative than as a data record. Major life-events are largely matters of emotion, and I would suggest from some experience that the more hard-headed amongst us sometimes fail to cope as well with such situations as the more emotionally-literate. While rationalism is of course useful, its tendency to devalue subjective experience is destructive to the quality of human lives. In educational terms, factual information only really becomes meaningful learning when it is mediated through the experience of a real human being.
Experience would suggest that such people also tend to see the management of institutions such as schools as a logistical exercise, rather than a human one. This might explain why they can appear insensitive to the disgruntlement they are wont to cause in their colleagues, and why they may make poor calls in critical situations such as recruitment, when an empathic ability to read character might be seen as an advantage.
If you are only relatively dimly aware of human sensitivities, it will be all the more difficult to factor-in the subjective elements that are needed for good logistical solutions. When planning a timetable, for example, does it really matter whether the patterns created cause difficulties or discomfort for the individuals concerned? I would argue that it can have a tangible impact on the quality of the resultant teaching.
And when it comes to lessons themselves, the same tendency is visible: we are neglecting the elusive, holistic experiences of human intellectual development in favour of a mechanised version of brain training. And we are preferring, as teachers, those who can deliver effective mechanical training over those who might have a more empathic, instinctive approach, who may value emotive quality of experience over clinical technical perfection. What’s more, those who are looking merely for technical ability will never understand the ‘something’ that many of the best teachers have, which largely comes down to one subjective, unprogrammable thing: charisma.
Perhaps those who think that A.I. will one day outwit humans are right after all – but it may be achieved not so much by building more powerful cyborgs, but by our own goal in dumbing down human life and turning it into a low-grade machine-like experience – which is what sometimes seems to be happening in parallel.