Box093: Are we calling AI ‘life’ now?

25.05.23

I was at the Creative Computing Institute on Monday doing the old melon-weighing thing while talking about our work at Arup. Folks, I’ve been going back and forth on this beard. I like the *idea* of a beard and people (Mrs Revell) say it looks good but I hate having it and I just see scratchy mess no matter what I do so, back to baby face. The pic seals it.

Folks, the new Sleep Token album is out and it’s been the only thing in my ears. I don’t know what else to tell you. The mysterious cultists have been trailing it for the last few months but it is a beautiful thing.

AI is life but life is not inevitable

‘AI is Life‘ suggests ASU astrobiologist and theoretician Sara Walker. (Since someone mentioned in passing ASU being at the centre of controversy-making I am starting to notice it more.) It’s a compelling argument that, in short, we should imagine all technology as products of life, in the same way that mineral sediments and oil deposits are. We don’t consider those ‘alive’ but recognise them as products of ‘life’ so; why not AI? It’s a compelling logic:

Many of us would not recognize mineral diversity as “life” any more than we would the computer screen or magazine you are reading this text on as “life,” but these are products of a sequence of evolutionary events enacted only on Earth. This is as true for a raven as it is for a large language model like ChatGPT. Both are products of several billion years of selective adaptation: Ravens wouldn’t exist without dinosaurs and the evolution of wings and feathers, and ChatGPT wouldn’t exist without the evolutionary divergence of the human lineage from apes, where humans went on to develop language.
Sara, Walker, AI is Life

I’m all for computation as a way of converging human experience with that of the planet: It’s the version of it I want to see, the extension of what Pohflepp described as the human ‘experiential register’ into scales to vast, tiny, slow and fast for humans to understand. Computers could and can translate the world of other beings to the scale to the human to make us more empathetic and comprehending, to prevent us projecting human standards on the rest, but this requires active effort from the activists, artists and scientists driving us to turn away from extractavist technology aimed at ‘efficiency and productivity.’

But, positioning ‘AI as life’ so simply serves the evil spirits of inevitabilism. The spectre of ‘Darwinian evolution,’ often misrepresented, has been used as cover for all sorts of nefarious and downright evil human activities because of the implicit argument that ‘if this is what the very building blocks of the universe have decided, then who are we as humans to intervene?’ Describing AI as part of an ‘evolutionary lineage’ (see also Romic on this) removes the agency humans have and need to express in order to direct it away from another technology used to wreck each other and our planet and places it in the domain of so-called ‘natural forces’ or ‘laws,’ arguments that have been used for generations to assuage guilt and make excuses for the worst excesses of capitalism.

On strictly typological terms, yes, there is an interesting debate to be had about what constitutes life and as folks like Bridle have argued, the imagining of artificial life opens up a re-assessment of what life is and how we relate to it that could inculcate better empathy and understanding of the life around us. I think Walker is trying to do the same, but the corollary or additional part to the article was missing. It should have said that we chose to make the Dodo extinct. We chose to decimate the rainforest. We chose to overfish the oceans. There was no inevitability to it, no evolutionary determinism, no natural law: Just as with nuclear annihilation, human cloning and the end of the space race, regenerative agriculture or fox hunting, humans have and had a choice over the life that’s left on Earth, and we have a choice over the one we chose for and with AI.

Upcoming

I’m taking part in a chat hosted by the Copenhagen Institute of Futures Studies on design and futures on 30th May. Look, they used the serious picture of me that everyone insists on using even though it looks nothing like me. (I think it’s from a phase where I wanted to look older and more ‘professional’)

Weirdly no one in this image is smiling even though we are all quite cheery people. Phil has dominated the shoulder score though. I was reminded last week that I recorded a podcast with Phil and Ben many moons ago that seems to have vanished. I’m not as bothered about this as when I spend days writing an essay only for it to vanish and the publication disappear but it’s still weird when something like that just ceases to exist.

Recents

I’ve been doing a lot of talk recently. I know said I was going to concentrate on Arup and the PhD but it feels like there’s still enough flex to go around and engage with people. I popped into Common Design Studio (which I helped start many years ago) to look at the student exhibition across Melbourne and London and offer some thoughts.

Reading

Finally got around to reading Ted Chiang’s ‘Will AI become the New Mckinsey?‘ which is his second very good bit of writing about AI.
On Joe‘s urging I sought out Design Epistemology. I’ve leant heavily on Cross’ definition of design as knowledge production for years; that it’s not art and science but (waves hands) something else. I usually supplement this with Frayling or Borgdorff’s notions of intuition and serendipity in discovery. The authors suggest that neither this nor the attempt to turn design into a science are sufficient epistemologies and either attempt to obfuscate design knowledge by keeping it esoteric (as in example A) or glorify and elevate science as the source of knowledge against others (as in B): “…design is hard enough without making it harder by applying esoteric theories inappropriately or by simplifying to such an extent that it is no longer functional or recognisable as design.”
One of the most interesting side-effects of the generative AI boom has been increased attention on the role of criticism in driving hype; that criticism can overstate the potentials of a technology and do the job of boosters for them. I first came across this in Nordmann & Rip’s Speculative Ethics work but I finally got around to Vinsel’s great Criti-Hype about the business model of hype-by-criticism. Although I guess his February 2021 claims of a new AI winter missed the mark, it’s a good piece to be mindful of when I catch myself re-quoting Google or McKinsey stats. (Also, pleased to see the tacit recognition that the car is the most dangerous technology in existence.)
Eryk Salvaggio’s prolific and brilliant blogging continues with In Defence of Human Senses.

Box092: Everything happens eventually

17.05.23

It’s Giro, which I managed to catch the first few stages of before travelling. Mixed blessings; I love it but it makes me very sad that I am still trapped off the bike and in front of screens.

I’m about the only person not playing Zelda this week. Instead I write to you from a very nice hotel in Charlottenburg in Berlin at the tail end of just over a week of travelling the Germanic world. I know I haven’t written in a while and I know I say that every time there is simply no good reason; I’m just a busy person. Every week I don’t write it is harder to again the next week.

I’m going back to the office more which means introducing two hours of sitting on a bus into each day (as opposed to 50 minutes on a bike #carssuck). As of yet I haven’t found a good way to blog or do anything with that time other than look out the window, listen to podcasts and do some Duolingo; I cannot type on phone keyboards and no one has solved the navigation problems of switching between windows effectively.

Recently, a couple of people were (jovially) saying I was really opinionated and it struck a discord somewhere. We can never know ourselves of course (I was recently told in an organisational psychology assessment that my results suggested I could be manipulative and I was like ‘how would I know?’) but I don’t know that it rings true. The things I have opinions about are often strong ones, usually because they’re developing theories, folk-observations, bewaring others of respect for water or about how much I hate cars. But there’s a huge amount of things (most sport, epidemiology, archaeology, theology, set theory, the Greek classics etc. etc.) that I have no opinion on, largely because I am reluctant to develop an opinion on something I have no understanding of. In the midst of Covid I was very clear that I had no opinion on ventilation, vaccination strategies or macroeconomic responses. Also I find the intellectual-progressive tendency to tear itself apart over its opinions really counter-productive/revolutionary.

Like, I understand the criticism of the AI letter, but isn’t the spirit of it right? Isn’t the spirit of the AI Dilemma right even if you disagree for critical or intellectual reasons with the individuals? ‘We’ expend so much energy on each other while the other side attack immigrants, ban abortion and gerrymander elections. I was actually sort of happy about the vibe of ambivalence around the coronation in the UK. I was expecting anti-monarchist vitriol from everywhere but happily people seemed to generally leave each other alone to respond how they wished and in a way that made them happy (other than the obvious nefarious arrests); whether that’s a protest, ignoring it or standing in the rain. If this weird spectacle makes you happy; great, people spend all sorts of money on all sorts of stupid things to make themselves happy.

Anyway, a lot has happened to tell you about though I can’t commit to giving you any more hot takes or elite opinions. They’ve all been used elsewhere.

PhD

Last week I capped off the chunk of work I need for the upgrade; the more-than-halfway point that is a submission of a good chunk of the work. Mostly this has involved addressing the front-stuff; tightening up the methodology and introductory sections but I also wrote a new piece on how I’m using theories from other domains. The process is really not straightforward; I loosely collect quotes or notes on things as I read in a blank document then iterate over and over and over again, grouping and cutting until an argument starts to emerge. Sometimes I’ll spend ten minutes trying to get a sentence just right or grab at an idea I can’t quite vocalise properly.

Anyway, it’s submitted now and when I get back tomorrow I’ll review the feedback from Matt, Joe and Wes, make any further changes and send off to the examiners. Feels good to be having it done but I know that really it’s just firing the starting gun on a much bigger chunk of work to finish the whole thing.

Recents

Can you believe it’s been almost six weeks? I’ll try and rattle through recent activities as quick as possible but in no particular order.

I was in Milan to do a panel with Viraj Joshi, Alexandra Mihai and Simone Rebaudengo hosted by Domus Academy and the Speculative Futures Milan guys on AI and design. Easily one of the most fun and interesting panels I’ve done in ages. We laughed about the fun stuff and got serious about the serious stuff, and had a wonderful evening rave in the hotel lobby. It was also great to be back in Milan for the first time in years and just run into loads of old friends, meet some new ones and talk about design again. I forget sometimes that people think about me as a designer and want to talk to me about design, no middle management.

Malina and I did a class on speculative design for Milan’s ‘Service Design Masterclass.’ I was also the inaugural guest on the Service Design Network’s YAP podcast talking about design, futures and working at Arup. Back on From Later again with Rob, Syd and Viraj (once again) talking about cloud seeding and CGI and all sorts of nonsense. Did a good research-heavy design talk for the Service Design College. I’m sort of glad no one’s asked me to write anything for a while. I can use my words good with my mouth brain but writing is hard.

Closing event last night of regenerative futures with Anna from Ellen MacArthur, Allison from Space10 and Martin from Arup’s circular economy services team. All brilliantly hosted by Franzi, the European foresight lead. Also when will Apple sort out this absolute travesty of a photo system. It’s easier for me to screenshot a photo in Photos then upload here than to go through the export process. Heic is killing culture.

Anyway, I’m in Berlin because we’re at the tail-end of Arup Foresight’s Regenerative Futures exhibition and events here for Berlin Design Week. The team made a new film – ‘Common Ground’ – and we ran a small series of public and private events which were bloody lovely. Before that (last week) I was in Austria for Desired Futures 6. Three days of great talks and discussions from a who’s-who of futures nerds. On Friday I hosted Julian in a conversation about imagination which was great.

Julian is massive but then there’s two of me so it evens out.

I’m sure there’s things I’ve missed and there’s lots of the same coming up tbh. I know there’s like three talks next week and it’s really hard to keep on top of it all. I do try and share them in good course on the social networks. My phone is buzzing to say I have to go elsewhere so I need to cut this short.

Short Stuff

Sorry, a lot of this is old news.

Not a huge amount but a bit of insight on tech companies laying off or redeploying ethics teams. They say; ‘we’re developing AI responsibly anyway’ we say; ‘inconvenient cost in race-to-the-bottom of the market’
Also Washington Post mapping out the different camps in the AI debate and, importantly, what their financial incentives are.
I’m very much on the ‘Robots have never and aren’t going to take your job (but they will take the things you love)’ side of things, Noah Smith’s analysis of an analysis of why 300 million jobs aren’t going to be erased.
Brazil is looking to get Chinese investment to start a semiconductor industry.
Emily Bender on tech wsu – ‘critics say’ as a positioning tool
Matt Webb and Campbell Orme’s prizewinning LARES. Matt’s done a writeup here.
There is no AI from Jaron Lanier on his usual optimistic drive. Unfortunately, he assumes that given a fairer system, everyone will play fair. For example, one suggestion is that people referenced in training data (eg. illustrators of cats) be paid for their effort when a new thing (cat image) is produced. This fails to take into account all the companies that would just spam training data (with cats) to overweight their contribution and profit. Over and over again, problems with AI are problems with capitalism not AI.
Side FX from K Allado-McDowell, really thoughtful reflection on the beginnings of critical practice with AI.
Britain really isn’t doing so well in the eyes of the world.

Ok I love you, and it is easier if I keep up with these things each week so I will immediately start preparing next week’s letter for you ok.

Box091: Please. Stahp.

05.04.23

DS081: Trying to replicate mycelium growth (I can’t remember why now). This scene was inspired by this shot which I thought was a closeup of a swimming pool and realised halfway through was actually a shiny floor. Live and learn.

I’m really sorry it’s been so long. There’s been a succession of illness, visiting relatives, work overspill and frankly, recovery which have kept me from you. I was in Copenhagen last week for my first work trip since Covid really and certainly with the injury and I’ve absolutely ruined my back with leaning on the stick, so now at least I’m trapped in an ergonomic desk chair with this glaring tab open. PhD is progressing nicely; methodology rewritten, chapter in review. It’s a little less frenetic than earlier because I’m waiting for feedback and making tweaks. I also finally tidied up my Are.na which I’m v proud of.

Rob Horning (✝ Real Life) wrote a piece about the ‘Paralogisms of AI.’ This is a new word for me (meaning fallacious reasoning) but something I’ve been thinking about a lot – how certain rhetorical devices re-occur in drawing attention to the spectacle of AI even as they sneak to dismiss it. This is well-documented in Nordmann and Rip’s idea of ‘speculative ethics:’ (which is different to the other speculative ethics which is ‘speculative’ as in ontological rather than ‘made-up’) They talk about how discussion (either positive or negative) of speculative, future ethical problems with a technology end up distracting from the present reality of the technology as well as the current ethical dilemmas (ChatGPT might put a load of people out of work but it is currently exploiting Kenyan labourers.) Horning writes a lot about some of the rhetorical myths and plays of computation to get humans to engage with them but I really haven’t seen an argument laid out for the logical loopholes people go through in talking about AI futures other than perhaps in AI Myths.

Some of these have been explored on recent Ezra Klein’s, particularly the ‘it’s inevitable, so we must build it.’ This recent one with Kelsey Piper lays out an interesting paradox though; that this attitude is not representative of AI culture more generally. The people who don’t think this line of AI development is a good idea and worry about the dangers, aren’t building it or advocating for it. This means there’s a selection problem in what we hear from AI people about what they think the future of the technology is; those who disagree aren’t being listened to.

I also want to setup a good AI thread. Somewhere where arguably good breakthroughs and applications of the tech are documented. Time was you would just set up a Twitter bot, but I’m not sure if that’s the way to do it nowadays. Maybe a project for a quieter time.

Recent / Upcoming

I was in Copenhagen last week on my first post-Covid, post-nine-smaller-femurs trip. I was there to help open up the Designing Narratives and Evoking Change course for Danish Design Centre which ends today I think. Also did some trips to see various friends around Copenhagen including Crystal who’s installing some grass at a museum (there’s more to it than that, as with all Crystal’s work it’s Smart.) I think it was an important talk. I started looking back on the What if Our World is Their Heaven stuff and realised that all the stuff I’ve been talking about there has moved from weak signals to IRL, which required a bit of a rethink.

Upcoming, I’m doing a talk for Service Design College on design research and critical thinking tomorrow (30 March) in the morning. I think it’s free.
Going back on From Later next week but I can’t think about what to talk about. All the signals are really loud.
I’ll be in Milan for Salone April 20-22. I’m doing a panel with some friends which I don’t think has been announced yet, but lmk if you’re about and want to hang out.
I’m delivering another masterclass for Milan Poli in May with Malina Dabrowska on speculative design (yeah) and design futures.

Short Stuff

Language is our Latent Space from Jon Evans is basically a way of redescrbiing Hayles’ cognitive assemblages as all the tech people imagine they are inventing STS for the first. time.
Brilliant review of Apple’s Extrapolations and the way it completely misses the mark on how fiction should talk about climate change (“It’s the fossil fuels, stupid”)
In an extension of the logic around Lil Miquela, Levis is going to use computer generated people of colour as model.
The reading list from Stochastic Parrots day.
Ed Zitron on the deflating of expectations and delivery in Silicon Valley.
Another John Lanchester great on silicon and Silicon Valley via Natalie.
The Crypto world is still all just crime.

This post was in drafts for ages and then I accidentally clicked ‘restore backup’ out of morbid curiosity and it deleted it all so there you go, lesson learned. Ok I love you love you love you.