AI and the enclosure of imagination

Artificial Intelligence could kill us.

The idea of machines developing intelligence, 'becoming sentient' is now such a well worn sci-fi trope as to have almost become a cliche. In '2001: A Space Odyssey' the spaceship's computer Hal9000 becomes self-aware and silently murders the sleeping crew. In 'The Terminator' the malign robot is Skynet, which starts a nuclear war in order to wipe out humanity. In 'The Matrix', The Architect plays a similar role. In all of these stories the machines become self-aware due to their huge computing capacity and unleash physical attacks against their creators, battles which they win by employing brute force.

Back in the real world, the actual danger to our societies from AI is not from possible material threats, but from an invisible, metaphysical undermining of humanity. It's not hard to imagine a future where our capacity for thought and understanding become hollowed out by reliance on AI, if we choose to delegate our thinking to the machines.

Thought is what makes us human and gives our existence meaning. Our ability to observe, comprehend and interpret the world - thinking, in other words - is at the root of what life essentially is. Without thought, we are nothing. And yet the recent development of advanced AI models challenges this most obvious and basic tenet of our selves.

The story of the human species is, largely, the history of the development and transmission of ideas, of the thoughts that spring from our imaginations. The harnessing of fire, the skill of cooking, knowledge of flora and fauna, technologies like the wheel, mapping the stars, mathematical reasoning, agriculture, writing, the concepts of nation and democracy, the understanding of diseases and of new ways to combat them. It is all of these, and innumerable others, that have allowed we wobbly-headed apes to dominate the whole global ecosystem.

Every single development in the long arc of humanity's rise has ultimately sprung from the brains, and thence from the intangible collective understanding, of the human population. But since about 2023, for the first time in AI we have an entity that proffers an alternative. Instead of having to do our own thinking, we can apparently outsource this task to something else.

In a fraction of a second, AI programs can produce apparently complex answers to virtually any question we put to them. They can 'write' entire university dissertations in the time it would take a human student to open a single text book. Alongside 'writing', AI can 'paint' pictures and 'compose' music, all by drawing on vast, digitised collections of human-authored books, articles, drawings, images and songs. AI-generated content is becoming increasingly difficult to distinguish from that which it seeks to imitate. In a few years, with more computing power and even more 'training material', being able to make this distinction will become all but impossible.

Thinking can be painful, and creating art is arduous. Why would anyone want to continue with the slow, analogue, error-prone process of human thought, if there exists an enticing shortcut? The problem is that people will start to ask themselves - indeed, people will already have started to ask themselves - what their entire purpose is, if they are no longer required to think for themselves. If we become reliant on AI, we will become alienated from the past, from one another and, most agonisingly, from ourselves.

AI is already having an insidious impact on the field of education. Students' creations, however grand or humble, are produced only for their own sake. School work comprises a series of exercises designed to hone the skill of thinking. The potential already exists for a school's programs of study to be designed and resourced by AI, with assessments created using AI, students' answers written with AI, and those answers, in turn, marked and commented on by teachers, using AI. We know instinctively that this scenario would be a charade. It would be a pastiche of education, a mockery of real thought.

But is AI not just the latest in a very long line of tools which transform our external realities, and thereby simultaneously our inner lives? We invented agriculture, and turned our backs on our hunter-gather existences. The scientific method gave us ability to transform our environments, and as a result it altered forever our relationships with the natural world. Perhaps AI is merely one of these ideas, these tools: transformative, but morally neutral?

Context and purpose, here, is everything. The big hitters in the AI world, OpenAI, Google and Meta among them, are private corporations. Their motivation for producing these models is to increase their product range and afterwards their market share of that new product. Free, limited access to the AI programs can be had for free, but fully specced versions are only available to those with costly monthly subscriptions, or as part of corporate licences. The product they are selling is our own capacity for thought, repackaged and flogged back to us.

In the coming years, as we learn to use AI more effectively and it becomes integrated more deeply into our lives, we will become ever-more reliant on it. Slowly, and then very suddenly, the societal habit of thinking will atrophy, like a muscle wasting away, and our sole choice will be to make even greater use of the AI that led us down that ruinous path. This is what is meant by the enclosure of imagination.

We have hundreds of years of experience of enclosure, and its damaging effects on society. In England, it was common land that was enclosed first: fields that had belonged to the village were fenced or hedged off for the singular use of the landowner. These land grabs were legitimised by acts of parliament and took away from the common people (literally) the means of their survival, their ability to feed themselves. Those people forced off the land had to migrate to newly industrialising towns where they earned wages, in order to pay for the food that they had been able to grow for themselves hitherto.

Since those first land enclosures in the 12th century, business has found many more commodities to acquire, marketise and sell for profit. There range from the more obvious privitisations of previously public utilities, such as the UK's water supply, to the less tangible ones. Facebook's business plan rests entirely on the enclosure of the idea of friendship, by first shifting the norms of social interaction, and then giving us targeted advertising while we try to reconnect with people it is suddenly harder to hang out with off line.

So AI could kill us; it could precipitate a vast moral and spiritual dislocation that undermines our way of life. On the other hand, given a concerted and concentrated effort, AI could be put to entirely different and more positive uses, though that would require a quite radical change in direction. As ever, whether that happens or not is down to political choice, and whether we have in it ourselves to recognise the problem, and think it through for ourselves.