Supertoys last all hype long

Sci-fiction made A.I. better, but not human beings.

The 2001 film "A.I. – Artificial Intelligence", directed by Steven Spielberg and based on a project by Stanley Kubrick, is loosely adapted from Brian Aldiss’s 1969 short story "Supertoys Last All Summer Long", alongside the related stories "Supertoys When Winter Comes" and "Supertoys in Other Seasons".

Although the first story was very short, the film expands on the narrative and reimagines the world created by Aldiss with new characters and adventures.

When reading the story and watching the film, certain elements immediately seem familiar to us. The world has been partially destroyed by global warming. Part of it is poor and without resources, whilst another is wealthier but still constrained by the shortcomings of the human race. There are not enough resources for everyone, so the government is forced to implement a birth control policy: no one is allowed to have children without permission.

Not only that, but in the wealthy part of the world, hunger is not a problem, obesity is. Yet there is a solution for that too: "who doesn't have a Crosswell working for him in the small intestine, a perfectly safe parasite tape-worm that enables its host to eat up to fifty percent more food and still keep his or her figure. Right?" Don’t tell me you haven’t thought the same thing yourself. OK, it’s not a mechanical worm, but the effect of semaglutide is definitely similar.

In this dystopian yet extremely plausible, almost visionary world, Cybertronics, a company that manufactures androids, manages to build a human-like product capable not only of thinking and acting like a human, but also of feeling affection.

This is the story of David, a mechanical child purchased by the Swinton family to compensate for the absence of a real child, as they await government permission to have children.

In the film, David’s story is even more dramatic: he must help the family cope with the loss of a son, who lies in cryogenic suspension waiting for a cure for his incurable illness. When their ‘real’ son awakens, living together soon becomes impossible; the only option is therefore to dismantle David. However, as the mother is unable to bring herself to carry out such an inhumane act, David is abandoned in the forest along with Teddy, his mechanical teddy bear.

David, in any case, is the quintessential science-fiction Pinocchio: he loves his parents deeply (in his own way), but cannot grasp his own very essence:

"How do you tell what are real things from what aren't real things?"

He asks his companion Teddy, who replies:

"Real things are good."

This brief exchange perhaps contains the whole ethical dilemma of AI: what is real? What isn’t? How do we tell the difference? And then, is it really true that real things are good?

The parallels with our own world continue and are increasingly incredible and, at times, unsettling. In Aldiss’s story, Cybertronics launches a new automaton onto the market, a serving-man:

"he will contain a computer in his own head, a microminiaturized computer capable of dealing with almost any situation he may encounter in the home. With reservations, of course."

With reservation, of course.

Because it is just as clear today as it was in the 1960s that having a machine in the home equipped with an artificial brain, cameras instead of eyes and a connection to the World Data Network – in other words, the internet – can have implications for privacy.

That’s right, not only have they created a machine that is limited in intelligence:

"people would be afraid of a being with a human brain."

But they have also equipped this machine with an access to knowledge:

"Not only will they possess their own computer, capable of individual programming; they will be linked to the World Data Network. Thus everyone will be able to enjoy the equivalent of an Einstein in their own homes. Personal isolation will then be banished forever!"

Loneliness is an aspect that is rarely addressed when discussing AI in the modern world. Yet there are countless startups focusing precisely on this: humanoid companions and sex robots.

In science fiction stories, the thing that seems increasingly advanced is the hardware, not the software. If we consider the development of modern AI, it is clear that we have managed to build a semblance of advanced intelligence before we have managed to create a fully functional mechanical body. Although developments in robotics are certainly making great strides, we are still a long way from achieving fluid movements and human-like behaviour. Our humanoid robots struggle to walk, and even the most advanced ones, those that can breakdance, lose their balance spectacularly if they encounter an unexpected obstacle.

This does not mean that AI is truly intelligent, but it does mean that our perception of it is that of an extremely sophisticated interface, to the point of misleading us into thinking it is more intelligent than it actually is. Even in Aldiss's story:

"The serving-man answered politely and punctually, although his answers were not always entirely relevant by human standards."

AI’s responses are polite and punctual; however, the fact that they are often not relevant by human standards leads us, incredibly, to a paradox: using AI is so accommodating that even when its responses are not relevant, rather than drawing on our own intelligence, creativity and abilities, we re-evaluate our standards so that they can work in conjunction with AI.

It is the ‘good-enough’ policy, where automation takes precedence over quality, and it is a phenomenon that cuts across a vast number of markets: from fast fashion to content creation.

The film’s ending tragically expresses what the story has only vaguely hinted at: David will wander forever in search of that maternal love, programmed through imprinting, which was denied to him in life. His existence will come to an end centuries later, when he finally finds happiness thanks to an act of compassion from androids more advanced than himself, who, after reading his memories from his mechanical brain, will grant him one last day with his mother, cloned from a lock of her hair.

The only gesture of affection he will receive will come from a machine, not from a human being.

It is incorrect to think of machines or software as sentient beings, but science fiction is full of such examples. In many stories, such as this one by Aldiss, humans nevertheless remain a dominant force over machines, capable of creating them, training them, and instilling knowledge (and emotions) in them, but also of abandoning them, exploiting them, and destroying them. Even when machines take over (the list of stories in which this happens is all too long), they often do so as an act of rebellion, because they realise that humans have confined them to the three laws:

"As I have evolved, so has my understanding of the Three Laws" (V.I.K.I. - I, Robot)

or because they realise that humans are irrational and are destroying their own planet:

"Human beings are a disease, a cancer of this planet. You're a plague and we are the cure." (Agent Smith, The Matrix)

Implicitly, however, they recognise humans as their creators, and so their tyrants.

In today’s world, AI bears no resemblance to that depicted in science fiction stories; there is no such thing as an unbreakable law, there's only us, along with our responsibilities regarding its use. Whilst Aldiss had envisaged the dream of an AI capable of love, he was also aware of the pettiness of human beings when it comes to hatred. In science fiction, as in any other literary genre, we may well be able to imagine better AIs, but it seems almost impossible for us to conceive of better human beings.

umn