Ultimate Things
That great philosopher Oscar Wilde said, ‘Most people are other people. Their thoughts are someone else’s opinions, their lives a mimicry, their passions a quotation’. The fact that when anyone interacts with (interacts with, note: not relates to) a chatbot, a GPT, an AI, and that these are aggregations of other people scraped by the Large Language Model in question, such that when prompted they proffer other’s opinions, mimicries of them, paraphrased quotations from them, the question arises: What’s the difference between a bot and Wilde’s ‘most people’? Are the former the latter, or the latter the former, or are they both two-dimensional mimics?
Part of the answer one wishes to give is in the parenthetical aside about ‘interacting’ not ‘relating’. At least so far, the most one does with a bot is interact with it (‘it’, note, not ‘him’, ‘her’ or ‘they’) because – again at least so far – there’s no one home in a bot. It might be a still-accumulating compromised entity just as any human is, but it does not have feelings, a biography, a limited but complicated history of relationships with parents, friends, lovers, spouses, children, colleagues, it does not have responsibilities arising from these connections, nor needs and desires relating to them. And it does not have an official individual identity recorded on some permutation of passport, driving license, bank account, mortgage, utility bill, insurance policy. In short, it is not a human person. (‘Human person’: since ‘person’ is a forensic, i.e. moral and legal, concept such that corporate bodies like the Shell Oil Company and the BBC are persons too, the qualification ‘human’ is required.) A bot is not a person in any sense, still less a human one; it does not itself have any moral and legal obligations and responsibilities, and is owed none by those who interact with it (which means it cannot make claims on them). It cannot be insulted or offended, cheated or deceived, libelled or defamed. Indeed it cannot be harmed in any way. But! – it can do harm. It can lie, mislead, ill-advise, even cause death. A strange asymmetry.
The legal persons (OpenAI, Anthropic etc.) who manufacture AI chatbots and offer them to our use (did we request them, by the way? want them, need them? just asking) can be harmed, sued, defamed, and the rest. Can they and should they be held accountable for any harm their bots do? The companies, qua legal persons, have a defence: that even they do not quite know how AI works and how LLMs respond to prompts. They themselves did not intend to do harm. The bots did it. Unlike the owner of a ‘pet’ lion which mauls a passer-by, they could not foresee and did not anticipate the harm that might result from how a LLM predicts what string of words to produce in response to a prompt. How could you sue Orville and Wilbur Wright for a plane crash today? Could you sue Smith and Wesson or Gluck because someone shot someone else with one of their guns?
Things will change if bots become conscious: self-aware, with personalities and intentions. Some claim that they already have become conscious. They point to examples of AIs taking action to avoid being shut down. If some Super-Turing Test established consciousness in an AI, would switching it off be a form of murder? Would switching off an AI that did harm be justified as a form of capital punishment for AIs? Or would we place AIs in the same zone of moral regard as, say, cows, which we slaughter so that we can eat them, giving greater weight to the pleasure of consuming a sirloin steak than to any suffering on the part of the cow?
The advent of social media in the noughties of this century has wrought a dramatic change in people’s behaviour. Almost everyone is glued to a screen, kept there deliberately by the algorithms, bombarded with opinions from others most of whom have no expertise to give opinions, unnerved by the shortcomings they find they have in comparison to the content-tsunami of claims, adjurations and examples of looks-maxing, manosphericism, prejudice, dogma, emphatic partisanships, anger, and quick-fix snake-oil advertising. Leave aside the wreckage caused to attention-spans, book reading, time for deeper thought, the flabbifying of intellectual muscles, and consider only the anxiety, the insecurity, the isolation, the manufactured ignorance that comes from being overwhelmed by unassessable ‘information’, the superficiality of swiping left or right and the nature of their outcomes – and more. To begin with, in the noughties, on early Facebook and My Space, it was other people (actual other people) making people ‘other people’, driving the already-existing Wildean ‘being-other-people’ to new heights. But now the bots are here this has ramped up so manyfold that the situation in the late noughties looks like a bow and arrow in comparison to an Anti-Satellite Laser Weapon.
Two striking anatomisations of all this were offered last week: Alex Ross on the late lamented Jurgen Habermas in the New Yorker:
https://www.newyorker.com/magazine/2026/06/22/jurgen-habermas-desperate-fight-for-democracy
and an absorbing podcast interview by Ezra Klein of the New York Times with Gary Shteyngart on the latter’s novel Super Sad True Love Story which, published in 2010, uncannily anticipated the situation today:
Alex Ross is powerfully eloquent and tellingly acerbic about ‘the barrage of toxic gibberish that constitutes the modern public sphere. Your e-mail is overrun with spam, scams, and smut. There are voice mails from no one about nothing. A glance at the news reveals that the President is continuing to spew lies and obscenities; that a trillionaire is peddling white-supremacist propaganda on a social-media platform he owns…Clickbait headlines…treat you like a starving rat in a Pavlovian experiment. A.I. systems simulate the experience of talking to an arrogant ten-year-old boy who knows far less than he thinks he does. When pressed, the chatbots admit that they cannot “naturally understand human morality, dignity, culture, or meaning.” It all adds up to a continuous discursive tinnitus—a buzz of random, fake, stupid, sinister chatter that nobody wants and nobody can stop.’ Compare that to Habermas on the public sphere and the kind and level of discourse needed there for the health of society: reason, and the Kantian ideal of Menschenwürde (human dignity). Add to it the insights of Adorno and we have, says Ross, the possibility of a dialectic moving ‘between crashing despair and hovering hope’.
Shteyngart imagined a situation in which automatic rankings of people – their looks, their credit ratings – are publicly broadcast to everyone’s hand-held smart device, so that when you walk into a bar everyone can know you’re rated the thirtieth ugliest person in the room, or as you walk down the street your credit rating is shown to the world; meanwhile, the US has a fascist President, ICE-like paramilitaries patrol the city streets, undesirables are deported, the US has invaded Venezuala (yes: that’s all in the novel, 2010) – and so for other scarily prescient details. It predicts life in the magnified glare of comparisons, without essential privacies; it predicts today’s dispensation, a life artificialized into being nervously or clamantly projected onto its unforgiving stage, and overwhelmed into ignorance and shallow emphatic emotion.
The Shteyngart interview ends, however, on a positive note, by urging what is still eminently possible: what he calls ‘sensualism’ but meaning by it something far more classically Epicurean than Cyrenaic in mode – that is, enjoyment of genuine humane pleasures, not mere crude indulgence in their chemically or thoughtlessly induced (still less screen-prompted) simulacra. Good food, delicious wine, in-person (‘In Real Life’, IRL) time with friends and those one loves, reading, reflecting, relaxing. Simple. All it requires is setting screens aside. The demons whirling and writhing in the dark universe of the screen are thereby shut away and silenced. The realisation that much of what they scream is ‘from no-one about nothing’ and that the algorithmic grappling-hooks they have sunk into your brain can be picked out without injury – indeed, with healing – makes the recovery of a Habermasian domain of civilised relationships possible: IRL.
This is what I meant by retaining and protecting the ‘essential self’ in my Challenge of the Future. For not just the aim but the evolving reality of the universe populated by nominal selves, the aggregated ‘other people’ society, is to imprison the essential human self in the dungeons below the screen-walled vaulted halls filled with the deafening tinnitus ‘of random, fake, stupid, sinister chatter’. We are trooping down thither, already in chains.
None of this is to say that there are not many positive uses of AI. Indeed there are. But there is a need to learn from history here. Think of our ancestors domesticating animals, taming the wild grains, irrigating the fields, in short mastering and adapting what hitherto was rampantly unmanageable – compare a rearing wild horse to its descendant in dressage and show jumping. Well: one might regret what went into such subjugations of nature’s native spirit, but you get the analogy for AI. It is beyond urgent that this Superman technology is harnessed to its manifold positive possibilities, and severely restrained from doing further and new – greater, perhaps existential – harms. At present the threat of its escape into unmanageability is being driven by the old, old Satan that poisons as it devours under the flashing neon lights of advertisements and blandishments: profit. Or if the Ultra High Net Worth Titans no longer care about how much more they make, because profit must now seem an irrelevance to them, they still have the lust to ‘Be First’, to win the race. The message therefore is: if they are not pulled back from this brink, their Being First will be everyone else’s Last.
Why is it that this amounts to cliché, this last remark - and yet nothing is being done? Are we just to go lemming-like over the cliff in a final spectacular burst of how smart we were as a species to create something so much smarter than ourselves, our last-ever invention, something that can write a suicide note on our behalf far more brilliant, eloquent, poetic than we ourselves could write - even though we seem to be trying to draft it now (with, doubtless, chatGPT’s help)?
This is not a rhetorical question.
Maybe it's just me as a senior man who has read widely, but AI fills me with unease.
This is a problem yet to be solved. Another problem is the 30% of overworked, underpaid workers who believe that Farage and Lowe are the solution to the problems they have in their lives. Where are the well paid jobs for a million 18-24 year old NEETS?