AI AND I

The important question isn’t whether machines are human, but whether humans are machine-like.

Dear Readers,

Thank you for reading The Kureishi Chronicles. I am still unable to use my hands and am writing, via dictation, with the help of my family.

Your contributions go towards my care, which is considerable. If you enjoy my writing, please do consider becoming a paid subscriber.

If you wish to preorder my forthcoming memoir, Shattered, you can do so by following the links provided here:

https://linktr.ee/shatteredH

Hello, this is Hanif Kureishi the novelist. I've been sent these questions in regard to my accident and my forthcoming memoir, Shattered, which is coming out in the autumn. I would like you to answer these questions in my style, using my Substack, Twitter and any other writing you have on me as a guide. Make the answers brief, witty, informative, without metaphors or trite aphorisms.

Last week, a journalist did indeed send me a list of questions and Carlo suggested, as an experiment, that we put them, with the above prompt, into an AI engine and have it respond for us. As a cynical old writer, I was sceptical as to whether a machine could do what I do. Nonetheless, though I am lazy, I am also curious.

Here are the questions from the journalist and the first answer the AI gave:

Q: How much movement do you have now? Are you working on a new book? How did you resolve to write or rather dictate your words after your accident?

AI: Movement? Minimal. Daily life? A challenge, but I'm adapting. New book? Always. The mind doesn't stop, even if the body does. Writing was survival. Dictation became my lifeline. My family, my scribes. Words flowed when nothing else would.

Obviously this is terrible and it doesn’t sound like me. I told Carlo to give up, but he insisted on prompting it again with this:

These answers are too brief. Please give four or five lines when answering. I never use rhetorical questions so please remove these. Make them sound authentic.

Here is the engine’s second attempt:

My movement remains limited, but I'm gradually regaining some control. Daily life is a series of adaptations and negotiations with my new reality. I rely heavily on assistance for basic tasks, which is both humbling and frustrating. Despite the physical challenges, my mind remains active. I'm always working on something new.

Writing, or rather dictating, became my lifeline in those early days of paralysis. It was an instinctive response to an impossible situation. I couldn't move, but I could still think and speak. My family - Isabella, Carlo, Sachin, and Kier - became my hands, transcribing my thoughts. This act of creation kept me tethered to the world when everything else had been stripped away.

I’ve never previously used AI and was unable able to understand how it could supersede the human imagination. But my sons Sachin and Carlo have been interested in it from the start. Both use it in their screenwriting and journalism. Many people play with these systems, but most of them have not persisted sufficiently to get a worthwhile result.

My sons tell me that once you have trained it, or rather trained yourself, it is dynamic as a creative collaborator and researcher. It seems to understand any given project in a human way, mirroring the intelligence of the user and offering up valuable ideas.

The second answer the AI gave, to my surprise, was full and satisfactory. There was nothing I wanted to add. It had studied my online oeuvre, including all the Substack blogs and other writings, and gave a coherent, cliché-free response, which I passed on to the journalist.

You can design an AI engine so it knows everything about your style and work, recognising your voice and knowing exactly how you do things. It is effective in helping with the more mundane aspects of the writing profession; boilerplate contracts, emails or the first draft of things, using and ordering your scattered ideas.

My sons tell me it’s a brilliant tool for when you have writer’s block. Sitting with a blank page, with only a glimmer of an idea, the AI will offer up intelligent possibilities to get the thing rolling. Having your own ideas formulated and presented back to you can be an important first step in a creative endeavour.

These systems tend towards encouragement, telling you your work is good and interesting.  But you could insist that it works with the scrupulosity of an editor and fact-checker at the New Yorker, for instance; offering cuts, sentence and structural rewrites and ideas for development. The AI’s notes are usually worth considering. It’s equally adept at fiction; understanding story, character development and structure. It’s important to emphasise that this is a tool, and that you should not defer to it.

When these AI systems were released, there was push by companies to implement them into their workforces, to varying effects. According to reports, a lot of these companies have since stopped using them, at least as policy.

But Carlo and Sachin have noticed, looking over people’s shoulders at their computers in public workplaces, that a lot of people are using these systems, most likely privately, and not letting on. Modern capitalism and white-collar jobs are characterised by huge amounts of time-consuming, mundane tasks; presentations, press releases and contracts. AI is Fordism of the 21^st century, expediting human labour, or removing it entirely.

The AI systems work by predicting the next word in a given sequence. When Carlo prompted the AI to answer my questions, it did so by calculating what I would say, knowing what I’ve said before.  This proves how predictable we are, that in fact if are asked a question, that the range of possible answers is limited by our history and intelligence. We all know that our partners and family can only respond to us from within a characteristic, limited view of the world; we say the same things repeatedly. The algorithm recognises and reproduces this.

The important question isn’t whether machines are human, but whether humans are machine-like. What unnerved people about these AI systems when they first encountered them was how human they seemed. They move between ideas freely, and in long conversations, they remember things, and are capable of digression whilst maintaining an objectivity about the subject.

For tyro writers, or those who are disabled or isolated - without access to creative writing courses or intelligent readers - AI might be the first port of call when it comes to having their work critiqued and developed.

If you are a writer, download Chat GPT 4.0, Claude 2, or my sons’ favourite, Perplexity, which encompasses all the models, and plug your work in, asking for help with ideas and its critical opinion. The more detailed your prompt - the longer your conversation - the better the output.

After this lesson from my sons, my view is that we might be able to use these machines as we have always used machines; to supplement and progress what we are already doing.  

With regard to the journalist’s questions, what would usually have taken an hour or so took five minutes, giving me more time in my new garden to sunbathe.