How to capture the significance of today’s discussion? There’s a lot I won’t touch on, like our fascinating conversation about error and creativity. I’m hoping this finds expression in my cohorts’ blogposts.
Our lecturer Hasan* began by sketching a map of our emerging knowledge. Thought this visualisation tagged with acronyms, we’re beginning to appreciate how these bodies of knowledge interrelate.
I like these sketches a lot – not only how Hasan makes them but also why. They’ve got me thinking about the value of diagrammatic drawing.
Hasan went on to explain he’d rewritten today’s lecture, which began with him showing us an image of a Rembrandt. We spent some time reflecting on our experience of encountering this artwork digitally. (Implicit here is our first-hand experience of these ‘masterpieces’ in galleries and museums as well the value they hold). We discussed what’s lost through online engagement (the encounter, appreciating the painting’s status as an object. We also spoke a lot about light. The difference between light reflecting off the painting’s surface and the light of our glowing screens).
And then Hasan told us this wasn’t actually by Rembrandt. The painting had instead been AI generated. How, wondered Hasan, does this knowledge change our relationship with/to the painting? We spoke about its context evaporating; its value lessening; there’s a loss – the word ‘horrible’ was used. This led to a fascinating conversation about originality and all the myths that attend it. So many paintings are co-produced but signed by a single author. TH made the point that restoration is also a creative process. It adds and changes what’s there.
So two forces are coextensive in this shift in understanding: There’s a sense of loss on the one hand and knowledge or information gain on the other. Hasan had rewritten his lecture to foreground this dynamic.
This conversation was designed to cast into doubt – or highlight our divergent and nascent views on – what creativity is and whether machines can be creative.
It also returns us to the old chestnut: What is art?
Other questions gathered, intriguing phenomenological, ontological and teleological ones about intelligence. What is it? What are its underlying structures? Can we even talk about it in the singular? Perhaps ‘intelligences’ make more sense. How do they interrelate? This was linked to Francois Chollet’s thinking about the interplay of general intelligence and task-specific skills. The latter is something that machines are very good at. The former? Not so much. To date, machines have little or no common sense.
After giving us thumbnail sketches of contemporary practitioners of techart doing interesting things, Hasan ended the session by quoting from Richard Freynman. The American physicist tells the story of talking to an artist friend about a flower (the nod toward Kant is undeniable)! Freynman takes umbrage at the artist’s argument that science does violence to matter by ripping it apart. He counters that what’s gained through this dissection is a greater appreciate for all the wondrous interconnections that compose life.
Hasan’s next and final move surprised me. He encouraged us to regard AI not as a blackbox but as the outcome of extensive labour. There’s so much to unpick here. It outstrips what I can consolidate in this post. Stay tuned for our collaborative project because it explores this idea by offering a portrait of a process through a little body of practice-led research.
I googled up the story of Freynman’s flower and found a recording of him telling it on YouTube. At around 02:14, he starts talking about the project of science. Its commitment to discovery, to finding out about the world. Freynman then goes on to advocate for living with ‘doubt’, ‘uncertainty’ and ‘not knowing’. Isn’t this a more interesting approach? Isn’t it wonderful that our answers might be wrong? This spawns all kinds of fascinating existential questions that bring us back to how we encounter not just The Next Rembrandt but all Rembrandts – all art. Everything. We just don’t know. And yet, we can understand and respond. This helps me to recognise the mystery of AI as a kind of fascinator. And it makes me think about the word ‘wonderful’. Could I be using it too often in some instances and not enough in others. AI is nothing if not full of wonder.
* ‘Hasan’ appears to be like ‘Cher’. They have no family name. NB ‘Hasan’ may be an alias to boot.