Published on Substack, October 2023

I WENT OUT TO DINNER THE OTHER NIGHT with three writer friends who I met at RMIT in 2000. We’re all still writers and I love them and when I see them, which isn’t often, I always get overstimulated, talk A LOT and it takes me a few days to calm down.

Anyway, that night, because we’re writers and it is 2023, we started talking about AI. AI, AI, AI. We talked about a lot of other things as well, but AI is the large scary black smudge that has appeared in the vision of writers and creators everywhere. It’s up there with catastrophic climate change, conspiracy theorists and neo-fascists. It’s a wonder we can see at all.

So, we had dinner and then as I drove home, I listened to an interview with Christian Madsbjerg from 2019 on the Time Sensitive podcast called “Why Design Thinking is Bogus”. It is a very good interview and I highly recommend it. Madsbjerg is Professor of Applied Humanities at the New School in New York. He also founded a company called Red Associates and I discovered him through his book Look: How to pay attention in a distracted world which is not a self-help book, but an exploration/guide to deep observation and how to see things with more empathy, accuracy and connection.  

In the “Why Design Thinking is Bogus” interview, Madsbjerg talks about the work he did early on in his career with LEGO. For many years LEGO was floundering and losing vast sums of money. They thought they were floundering because kids had shorter attention spans and played video games but when Red Associates sat down and observed children throughout Europe, they found that almost all the kids, even those with ADHD, wanted to do difficult things and that they liked complexity and learning. LEGO’s underlying assumption, that modern kids would only do things that were quick and easy, wasn’t true. As soon as LEGO began to make their designs more complex again, their fortunes improved.

Madsbjerg goes on to outline the approach he uses:

1. What are the underlying assumptions?
   (Modern children have terrible attention spans and they only want to do easy things)

2. Slow down.

   (Get down on your hands and knees and actually play with the children)

3. Look for counter examples.

   (For this he uses the example of the Mexico wall. What does it feel like to want to build a wall between Texas and Mexico? Or, what if children liked complexity and doing hard things?)

Madsbjerg argues that you cannot believe something to be true until you have properly explored the opposite.

When I woke at 5.00am the morning after meeting my friends for dinner, over-stimulated and obsessing about AI, I started to think of counter examples for why the people inventing AI could possibly want to use it to write fiction. Is it because they are not able to optimise and interfere with the experience of reading and they would really like to? Do they see the creative output of humans as meaningless? Are they trying to cash in? Writing is not exactly expensive and it’s certainly not like they pay writers much of anything. The conclusion, I ultimately came to was that if they can get AI to successfully write fiction, it will be the most human-like that it can be. It will be able to do something complex, meaningful and generative. This my friends, is the Turing Test and if you’re a tech-head as opposed to a paranoid writer, it’s bleedingly obvious.

I have always been a daydreamer and when I daydream, my default mode network returns me to the small island where I grew up: often dark scrubby forest but also, sometimes, the rocks beside Allports Beach. It is hard to describe what this is, it’s not memory, more like a sensory combination of experience and feeling. It is never linear, it exists inside me as a shimmering internal corporeality that I pass through, or drift across. I can’t know what another person’s internal landscape is, or if they notice it or care or dwell there, because even if I see them daily and know them intimately, it is inherently unknowable. And, in my mind, that unknowableness is why art, music and writing exist. Sometimes when you read a book, or see a beautiful film or piece of art, or listen to music, you catch a glimpse of it: the shimmering internal corporeality of someone else.

When I was a child and the weather was hot my granny or my mum would often take us to Allports. I remember that irritated, dull, hot feeling, and the supreme frustration of being stuck and dependent that drives some children nuts, and then the utter joy, the freedom of being wet and salty on the rocks. After we had swum and sun-baked and run up and down the beach, we would gravitate to the rock pools. It feels to me now, as an adult, that we spent hours up there. The hot granite on the soles of my feet. Plunging hands into warm rock pools, the slight schism of fear that tingled the tips of my fingers as my hand touched Neptune’s necklace, gelatinous solid little bobbles of weed that could also be a blue-ringed octopus, because my parents were always warning us that they hid there, in amongst the weed. Limpets that gripped hard to the rocks when you bumped them. And time. Time that stretched and shrunk. Time that twanged. Five minutes spent on the rocks as a child is entirely different to five minutes spent on a device. Or rushing. But in my body I have an understanding of those different types of time and what they mean.

Madsbjerg argues that because we now see ourselves as optimisable and machine-like, we consistently overlook the lived experience. And this is almost always to our detriment.

Everything we do as humans, everything we create, and each piece of research or knowledge we generate is triangulated with lived experience. For what we create to be authentic, we must tell the truth about our experience. In just about every writing course I have done, it has been drummed into me over and over: your writing must tell the truth. Observe. Watch people. Record what is really happening now.

My partner, who is a complex systems scientist and deals in data sets and algorithms and is often dubious about my AI claims, tells me that AI cannot critically describe the learning process it undertook to arrive at something. He says that even if it could, it would be so entirely different to how a human learns, it could only mimic human description of how it learns, as opposed to critically understand it in a human sense. AI cannot evaluate like a human because it has no good representation of reality to test against, it has no real experience. AI is entirely limited by its data. It cannot investigate and it cannot go out and collect more data to fill a gap in its knowledge, at least not yet. If there is a gap, then it must co-exist with that gap. Also, because AI relies on data sets, it will always be about the past. Never now. And for this reason, we must always be sceptical about what we get from AI.

If we think about Madsbjerg and his counter examples, it seems that no one is thinking about what it will be like to consume the material that AI is going to produce for us. They are thinking about optimisation and targets and costs and messaging. They are not thinking about us as people. I would love to read a book written by AI if it was about the lived experience of being AI, but I am not even sure that AI could write about the lived experience of being AI because it is so different to human lived experience, there is no language for it. It cannot know. It can only summarise and synthesise the descriptions and observations it has consumed of what it is like to be a living, breathing human. I guess this kind of cheap imitation of living is what the gatekeepers think they can get away with, but as is increasingly being reported, people are not complete fools. They understand the subterfuge and they turn off.

Corporations and late-era capitalists would like to convince us that we are optimisable, quantifiable, and machine-like but we already know what we are, in our lumpy bodies trying to get by on our devastating and beautiful planet, coexisting with our shimmering internal corporeality. All of us know what it is to jam a finger in the car door or to taste very bitter, sweet, ice-cold lemon gelato on a hot day. We understand it with our bodies and our senses, we use it to make meaning and we share this meaning with other people.