Thinking Around Artificial Intelligence
There are so many words around at the moment about Artificial Intelligence that I am a bit reluctant to add my own contribution, although letting it all go by doesn’t help either. The other day I got yet another email praising one of my novels (Book 2 of my Phenomenological Detective series – The Heights of Abraham) up to the skies:
“My name is Kate Walsh, and I’m a Senior Commissioning Editor at Macmillan.
I recently came across The Heights of Abraham, and I was intrigued by its distinctive approach to the mystery genre. Blending a compelling investigation with historical echoes, psychological depth, and philosophical exploration creates a fascinating reading experience. Patrice Lanier’s unique investigative style, combined with the Romanian setting, buried secrets, and the tension surrounding the unsolved murders, offers a refreshing perspective for mystery readers seeking something beyond a traditional crime story.”
So far, so AI. I have had loads of these, where my novels have been scraped in order to make me think the correspondent has read them. True, they do not usually say they are from one of the big publishers and do not carry the name of an actual commissioning editor who works there, their logo, and real address. The next paragraph reads:
“It made me curious to learn more about your work. Are you currently writing or developing any new manuscripts or future publishing projects? I’d also be interested to know whether you are represented by a literary agent. If you have representation, I’d be pleased to continue any conversation through your agent. If not, I’d be happy to explain how the process typically works should there be mutual interest in discussing future publishing opportunities. I would be delighted to hear from you whenever you have the opportunity.
Warm regards,
Kate Walsh
This is a bit different, mentioning an agent (everyone knows Macmillan does not entertain unsolicited manuscripts or non-agented authors). It seems slightly less obvious than most approaches – one normally realises immediately that any contact from you will result in irritating nagging and requests for money.
So I decided, for the first time, to contact Pan Macmillan (they have a special email department for this) and ask them whether it is genuine or not. I was not really surprised to find that it was not – although I did have a slight thought that it might be. Not much disappointment, though – and the knowledge that these AI nuisances are getting better at it.
I have written previously about the unpleasant affects this sort of scam can have on writers. A lot (most?) of us have variable confidence in our talent. Some days we are sure we are good, some days we want to throw the Work In Progress out of the window, followed by what we wrote it on.
Awful, then, to have to deny that what our many scammy correspondents write is correct; it feels as if we were denying that we really are a good writer whose work creates, ‘a fascinating reading experience etc.,’. We can only comfort ourselves with the truth that we did not fall for the unkind scam.
Additionally, we have to accept that our books are being scraped to train AI, without our permission. This is not, apparently, an infringement of copyright. Although it may be, in some way, an offence against Intellectual Property. On a bad day, it appears as if the word sc(rape) is well adopted.
Just as an infinite number of monkeys, with infinite time and infinite typewriters (because using word processors doesn’t affect the outcome that much) could produce the entire works of Shakespeare, the monkeys will still not understand any more than AI does.
***
The hot topic for writers, editors, publishers, illustrators, and fine artists, at present, is Artificial Intelligence – and how it might make us all redundant. Even the Pope has included it in his first encyclical, ‘Magnifica Humanitas – on safeguarding the human person in the time of artificial intelligence’.[1]
It makes little sense to dismiss AI, as some have done, something which will have its searing day and disappear into obscurity. Neither is it sensible to embrace it lovingly and adopt it wholesale with all its faults, as a way of doing less work and making more money. Or as a personality, with whom one can have a relationship.
There have been many articles written about how AI will destroy the world and many about how it will make it so much better. It clearly requires a deeper dive, a longform consideration, to alleviate some of the terror.
We should be aware that the apocalyptic thinking around AI could yet be a promotional gimmick – a way of forcing us all to think too much about it, yet another way of destabilising our poor fragile world in its weakened state. We need to keep
that in mind.
When I was writing my (not yet published) book about therapeutic care for people with dementia, ‘Making a Masterpiece of Our Time Together’, I compared music generated by a human composer with that generated by the Star Trek android, Mr Data, who desperately wanted to be human. It was clear that, although Data’s music was technically correct, it had no human sensibility behind it: it had no soul. In Star Trek: the next generation, Mr Data is sentient and treated like a human being with moral agency. Where he is not so treated, he takes the place of any minority group – and we are horrified at his diminishment.
The piece is dated now, but there is something which may contribute to our endless discussions about AI. Experts: techno-oligarchs, journalists, and Richard Dawkins, have parachuted in, with their benign/nasty agendas, with their opinions, well thought-out or not. What should we get out of all this noise? How shall we see the wood for the trees?
It’s not very good but it’s getting better: Many of us think we can detect work done using AI, and it is often true we can. Writers frequently get unsolicited emails telling them that their book is a ‘heartbreaking work of staggering genius’[2], which is the start of a scam to extract money from them to promote it. I have blogged about this previously, pointing out the unpleasant feeling authors have when they have to deny that the words the AI whispers about the quality of their work are true. We all need confidence boosts …we want to believe them. But we have to put their email in the recycling. Because what they write is there to scam us into giving them some of our extremely hard-earned cash. So, we can often tell now. But, probably, in the future, it will not be so easy. They will get better at fooling us.
Dangerous AI – Kill all the humans:
If super advanced AI is given a brief that it must defend itself, or something else, at all costs, it would then be clear that, if humans are perceived as an existential threat, they will kill all the humans. The answer, I suppose, is to make sure that doesn’t happen. Didn’t Isaac Asimov write a prime directive saying that no robot is ever allowed to kill a human? I looked it up:
The Three Laws, presented to be from the fictional “Handbook of Robotics, 56th Edition, 2058 A.D.”, are:
- A robot may not injure a human being or, through inaction, allow a human being to come to harm.
- A robot must obey the orders given it by human beings except where such orders would conflict with the First Law.
- A robot must protect its own existence as long as such protection does not conflict with the First or Second Law.
That’s what we need. Or some other procedure so that AI simply is not allowed to kill all the humans. Or are people who say there is this level of danger just spoofing us? Scaring us into thinking this is a problem at all? We don’t really know – in this era of lies, falsehoods, untruths, and more untruths.
Conscious AI – What Richard Dawkins thinks:
Because I have long had an issue with Richard Dawkins’s arrogance and poisonous obsession with religion, though appreciating his way of expressing himself, my interest was engaged when I read his opinion piece on AI being conscious. I am not going to analyse this – you can get that elsewhere – all I can ask is when he actually lost his mind? I was a bit stunned at his ‘christening’ his Claude AI ‘Claudia’ and making it his friend for life. ‘Christening’? Dear Richard. AI is NOT conscious; that only makes sense if you think that being conscious consists of being really good at guessing which word is statistically most likely to come after the previous one. AI is not your friend; it is certainly not your girlfriend. It only simulates understanding. Start looking for good people in your life; have a human conversation with them. You are not the only clever bloke out there.
The Benign Face of AI:
They say that it can make our lives (as creatives) easier, because it can scrape faster, look things up faster, sort and re-assort faster. At present, it certainly can do these things. But can it do them well? Just now, only with a lot of human help. Which may be more work than humans would have done, on their own, in the first place. Chances are that it won’t be able to reduce the amount of human work put in for a long time if ever.
Platform refusing Independent Books:
- Rakuten Kobo, one of the book sales platforms everyone uses, says that, in 2025, it rejected 45% of books submitted. They say that 80% of these are suspected of being produced using AI. Huge numbers, and it is hard to imagine so many AI books being submitted; not to mention how they are detecting the suspect ones.
- Amazon’s response: At least there has been one – or rather two – responses to the increases in AI slop from the Big River. Amazon asks authors to declare when publishing, whether they have used AI in the production of their book; and it has reduced the number of books which may be uploaded in one day to three (from 100). I am unsure whether the first will be all that useful (we could lie), and the second seems unlikely to affect slop-resisting authors.
- Hatchette and ‘Shy Girl’ – Mia Ballard’s debut novel, ‘Shy Girl’, was withdrawn by Hachette Publishers and, although it is still shown on Amazon’s website, it is not available and it is not known whether it ever will be. It was withdrawn because the publisher detected (presumably using the imperfect guidance available) that AI had been used in its production. The author denied that she had used AI but admitted that an early editor may have used it. The latter made no difference whatever.
Types of AI:
- AI Generated: All the main platforms refuse to publish books generated by AI. If they are aware of them and, as above, the detection software must be somewhat prone to errors.
- AI Assisted: Assisted is allowed – although I cannot be sure of how they decide what is assistance and what is not.
- Other uses (presumably assistance?): Brainstorming, sounding board, suggestions for improvement, first level title generation
Editing and translating: The question would be: ‘Are editing and translating mechanical functions’? Leaving both to AI would remove any agenda that editors and translators have. AI could probably remove duplicate words, spelling mistakes and typos (if it was made aware of the differences between UK and US spelling for example), timeline issues, errors in point-of-view (sometimes), but it would not understand nuance, unreliable narrators, extended metaphor. In short, it has not lived; it has no moral agency.
As far as translation is concerned, it is hard to imagine AI in the rich pastureland of international work, the culture and variety of colloquialism, the preservation of the author’s voice. The reproduction of these things usually comes down to the lowest common denominator, bland mush. It must be said that a book of 100,000 words, in translation, needs 100,000 decisions …
Covers: It is not just our prose which is being scraped, it is also our images. We have all heard of the AIs which took our photos of human beings and removed their clothes, the ones which put us in compromising positions. It works for book covers too.
Research: I can see that there may be some help here for research, because some authors, I believe, don’t do that much. Some look stuff up – and that is fine as far as it goes. AI can certainly help with it, much as Google does/did. It can give you a quick list of resources with little trouble or effort. You can just follow them. It may not, however, give you the resources you are begging for – the paper or book that only you can recognise as useful when you see it. Doing proper research is a job both specialised and specific. The AI never took a course on research methodology and materials, did it? Scraping a book is not sufficient on its own.
Then we have the problem of ‘hallucinations’. I thoroughly object to the word (as a longtime researcher into actual hallucinations and delusions of people). Sometimes, when AI doesn’t know something, it makes stuff up. Hence the increasing appearance of scientific and political papers which do not existand reports of research or legal cases which never took place. These are not hallucinations, they are errors.
Where do you get your ideas anyway? The answer to this is ‘everywhere’ – and there seems no sensible reason to exempt AI. If it has extra ideas to your own brain, your own reading, the scrap of conversation you heard on the street, why not? As long as you don’t let it do the thing for you – just use it for inspiration, with other things.
AI cannot think for itself because it has only been trained what to do – that is, it can only copy what has been done before. Creativity consists of many different things and different processes; it needs to jump between a large number of things – which is why authors, and other creatives, must read widely, look at other people’s art, talk to other creators, talk to other people in general to glean inspiration from the widest variety of sources. In human creativity, sometimes things which have never been combined before are brought together, making a new thing which has not previously been done or made.
This is a regular and strong source of inspiration – and I can see why the bringing together might be done by AI. It’s like that idea about writing a mystery novel by dealing a series of cards – infinitely(?) changing the story by shuffling them and dealing in a new order. But it is not really infinite, it only has a limited number of ways to arrange them – even allowing for random elements – although the scenarios may be quite large, they literally cannot produce something entirely new.
Environmental: The leading concerns, environmentally, are that AI needs a great deal of water and a great deal of energy, and this is very alarming in a world which is subject to galloping climate change. All our future wars and other conflicts (special military exercises?) are likely to be increasingly about water. We cannot afford this.
There is a difficulty, though, in that no one seems to know what sort of amounts of water and electricity we are talking about – mainly because big tech is not telling us. There are some figures, for example, for all the functions at existing sites, but not separating out those spent on AI. So everyone has to guess.
The search and storage requirements of data centres is also extremely rough on computing hardware. Graphics Processing Units (GPUs), burn out at an alarming rate – and as well as being stuffed with rare minerals, producing them takes huge amounts of water and energy and leaves behind mountains of toxic sludge.
There are additional environmental concerns, like noise, like pollution, and risks to natural diversity – these data centres will be very large indeed and cover a lot of natural countryside.
But, back to what creatives are worried about with AI?
- Being replaced? That is a serious subject, none of us want to be in the position of losing our jobs because robots (or similar) can do it better, faster, and cheaper. The existential fear remains in some form even though we are ‘sure’ that AI can never be as creative and original as we can. What is also likely to be important is that it is difficult enough getting seen in an overcrowded book market. If huge amounts of ‘AI Slop’ are being submitted, it has the potential to be many times harder.
- Scraping: The other big issue is the phenomenon of ‘scraping’. This has been said to be the greatest theft of intellectual property in human history. It is plagiarism, pure and simple and, the whole system is based on theft. Many of us have, all of us will, in due course, had our works sc(raped). No one has asked permission or negotiated a contract to use authors’ work to train Artificial Intelligence. This is infringement of intellectual property. It is also violation of our souls against our will.
Conclusion:
Probably use it if it helps but don’t let it get away from you. Don’t let it de-skill you, don’t let it make you lazy. Certainly don’t let it tell you about how to live your life, and do not let it advise you on relationships. Do I really have to say that? How ridiculous. We humans are quite good at messing up our own relationships without outside help.
What, then, does AI lack? What will it always lack? It doesn’t (and can never) fully understand the world, it has no moral agency, it lacks human vulnerability which nurtures nuance. AI has never lived.
And, by the way, don’t let AI shop for you. And don’t ever let it have your credit card.
Elizabeth Mostyn, 24.07.26
Further reading
Blood in the Machine: The Origins of the Rebellion Against Big Tech by Brian Merchant (Little, Brown US)
Enshittification: Why Everything Suddenly Got Worse and What To Do About It by Cory Doctorow (Verso)
The Reverse Centaur’s Guide to Life After AI, Cory Doctorow (Verso)
Technofeudalism: What Killed Capitalism by Yanis Varoufakis (Bodley Head)
The AI Con: Emily M Bender and Alex Hann
[1] It’s available in full on the internet, it’s 40,000 words. Try reading Part 3.
[2] This is from writer and coach, Randy Ingermanson. Thanks!
