An illustrators thoughts on AI

On William Morris, killer robots… and the “Final Frontier” of exploitative labour

As I type this, straining my carpal tunnel-plagued dominant hand, I’m seriously considering the longevity of my drawing career. For writing, at least there’s dictation, and for illustrating, well, let’s say hypothetically, I could just tell a computer what I wanted, and then… oh wait. That’s already happening.

Of course, in my version of machine learning models, the only input would be my own, including an archive of my previous works, mark-making, and techniques that I have taught it. I know that a benevolent kind of AI model is what most people want, ultimately, and yet I feel forced to pick a side: either you shed all notion that you could somehow be a Luddite, and welcome the new technology with open arms, or you join the outrage campaign against the inevitability of innovation.

As other artists do, I feel the same fear that the devastation of AI could bring upon an industry that we’ve built by our own hands. And yet, I hate the usual dichotomy in which we draw comparisons between AI and humans; AI – cold, technologically superior, feelingless; human – flawed, visceral, and mortal. It’s clear who gains the moral upper hand in this framing, by mere virtue of us refusing to imagine further than that (how boring). And I hate being put in the position of picking a side, because in all truth, I love Data – data with a capital D, that is, the lovable android from Star Trek.

Star Trek: The Next Generation, Season 2, Episode 9: The Measure of a Man (1989, direct by Robert Scheerer, written by Melinda M. Snodgrass)

There’s a compelling episode on the second season of Stark Trek: The Next Generation, called The Measure of a Man, where Lt. Commander Data’s right to consent, and thus his sentience, is debated in a hearing. Captain Jean-Luc Picard, who represents Data, argues that, if an android can fulfil two out of the three requirements necessary for sentience (intelligence and self-awareness – both of which Data clearly has), what measure of conscience is required to determine if a “life-form” is sentient – and how exactly, would conscience be measured?

No, I’m not suggesting that we go hand out citizenship to our MacBooks – especially not when so many human minority rights are still at stake. I suppose I see more of an opportunity to incorporate AI as a tool in our workflow, given the data is fairly obtained. I do not believe that humanity as a whole is ready to dispose of human craftsmanship entirely.

Indya Moore gives an interview at Essence black women in Hollywood event. Adapted from a video, via Indya Moore on Instagram.

Furthermore, though this may seem a little farfetched: if AI is “continually learning” – as they say – and moving more and more in the direction of becoming more sentient, I’d rather be welcoming than discriminatory to these new lifeforms; after all, they didn’t ask to be “brought to life.” Let’s just hope that future generations of AI are more like little Datas, and less like killer robots.

I’ll go on record to state that artists tend to be optimistic when it comes to emerging tech (after all, many of the tools we work are precisely that) – if you think on the other hand, that they are being hysterical about machine learning models possibly replacing them, then I invite you to think about how you would feel if, no, when, automation starts taking over your livelihoods.

It’s not an unrealistic thought – plenty of service providers have already been automated to a large extent, if not completely, without anyone batting an eye: bank tellers (ATMs), cashiers (self-checkout lines), customer service chatbots, website builders, spellcheckers, etc – what’s stopping the automation of everything else?

You’re yoga teacher? I don't see anything preventing the creation of digital avatars, modelled off of thousands of online content, replacing yogis who have been honing their skills for years. You're an accountant? Why don't we, from now on, just feed all company paperwork through a machine that can crunch the numbers for us! What about AI predicting the stock market? Bye-bye investment analysts!

Automation will happen across our entire economy, it is simply a question of when. I think we can say this with confidence, even while projecting exponentially into the future can be pretty futile, I mean, look at this thing:

A promotional illustration from 1899/1900 depicting what 2000 would be like, via Wiki Commons.

… but while we’re looking at the past, well, let’s look at the past: innovation hasn’t always spelt doom and gloom for the artist’s existence.

Take the Industrial Revolution, for instance, which brought forth William Morris and the Arts and Crafts movement. Photography, which was said to annihilate artistry in the 19th century, ended up being a vital tool to many impressionist painters.

Though it’s comforting to think about how Morris and his band of socialist craft bros may have influenced the tide of British tastes, turning attention away from machine and rote assembly to workmanship – what happened to skilled workers, e.g. knitters and weavers, who did not come from well-to-do families, and therefore did not have the luxury of an artisan getaway?

Enter: disciples of the mythical figurehead, “General Ludd,” the namesake for the now rather disdainful term “Luddite.” Though they weren’t completely opposed to innovation brought about by industrialisation (sound familiar?), they protested against the exploitation of labour and the complete evisceration of their established professions by smashing machinery to smithereens – only to be met with brutal and harsh government repression.

“The Leader of the Luddites” by an unknown artist, 1812. Yes they wore dresses as a form of protest. – can we call it drag?? Image via Wiki Commons.

In other words, it’s all swell to envision a future with Morris-esque arts and crafts houses, but under this organisation of the economy, lower-income creatives may have no choice but to be forced into “useless toil” – as Morris called it – or to revolt, without any certain prospect in their lifetimes of consumer habits turning in their favour. It’s completely understandable why people who have built their entire existence on their craft or a dedication of some kind – be it an artist, a musician, accountant, whatever – would dread the technological eclipse. All the while begrudgingly acknowledging that it is precisely these advancements that, to this day, have ensured our quality of life, complete with the conveniences of IKEA furniture, drawing tablets, iPhones, affordable clothing and the like.

We have to accept that we may not all be safe. It’s exceptionally cruel to be met with the flippant adage simply to “adapt to the times” – but yeah, we may have to adapt. Even with hope, and in the face of the most adamant, the sun always sets. And yet, in Lindsay Varty’s project Sunset Survivors, an interview with a steelyard scales maker reveals a particularly moving motive for defiance against modernity and even old age: her craft is the only memory left of her father.

This brings us to an idea that’s currently making the rounds in the AI/art debate – that human stories/feelings matter, and indeed the bias for human-made art is strong, as a study on aesthetic value by the University College London and the University of Copenhagen exemplifies. It is set by a moral and philosophical precedent, too, for example in Kant’s imperative to respect the dignity of human beings, to treat them as an “end” and “not only as a means.” Further, it is an idea mirrored by self-proclaimed Marxist William Morris’s precept for work that is worthwhile doing, namely that it is work containing three hopes: "hope of rest, hope of product,” and a “hope of pleasure in the work itself.”

Meme by yours truly, image via Wiki Commons.

Work that isn’t worthwhile doing, on the other hand, is one that forces lower classes into a kind of slavery which cheaply mass-produces not “wealth” but “waste” for the upper and middle classes. I may not share Morris’s particular taste for the artisanal, but I certainly agree that cheap and meaningless labour tends to be highly exploitative of workers and natural resources, resulting in more waste and shit that we do not need. Surely, the counter to many of these problems would simply be to allow people to do dignified and fulfilling work, and to reward them fairly for it? Doesn’t the inherit value of work and its products lie in the “hope” that people have in them?

And hope is just one of many intrinsic values that differentiate us from AI – it goes further than that. AI models do not… take part in society. They do not have relationships, argue about politics, or think about that time mum very awkwardly gave you “the talk” as you were trapped in the passenger seat beside her. And because they aren’t plagued by the same flaws, self doubt, carpal tunnel – maybe even pride – the way that people are, they copy too faithfully the stylistic prompts given to them, making it much harder to toe the line of copyright – heck, even humans struggle with this!

Simply put, AI models can never amalgamate inspiration and life experiences the same way that humans do, and all of these factors remain as intrinsic values to the art we make. Human stories matter. So there we have it: if AI can never dream or hope then we’ve won half the battle – right?

Speculation: what’s stopping programmers – even with legal checks to prevent copyrights infringement – from building a model so complex, with so many inputs and simulated neurons, that it gets as close to the real thing as you can get? What’s stopping future patrons of the arts from “hiring” AI models – because it will definitely be cheaper and faster for what they perceive to be pretty much the same quality – over us? What is stopping programmers from birthing a self-aware model with its own “hope” and entitlement to an origin story? What then?

And, conversely, what if, instead of consciously obeying typed-out orders, the AI models, to the dismay of their programming parents, choose an autonomous line of work? Could they lead an uprising against the servitude of man by man – I mean, man by machine?

I understand that this is all incredibly far from reality – unless this thoroughly creepy dialogue between a Google chatbot and a former employee has already convinced you otherwise – so I hope this may be excused for its necessity in excavating meaningful action, at least for me. I understand, also, that action is demanded over introspection, so I echo the demands of those already advocating for the cause: unquestioning faith in innovation threatens to displace all workers, and should always ring alarm bells. We must tread with utmost caution and care in face of these threats, or else start chucking Aldous Huxley’s entire bibliography in the bin.

The first threat that unchecked, rapid development of AI could pose is the one I previously mentioned: that of waste and repletion of our natural resources. We already know that industrialising manual processes leads to an increase of production, furthermore, according to a paper by researchers at the University of Massachusetts, training a single AI model can emit five times the carbon dioxide of the “average American car.”

The second set of threats, which I will briefly group together here but are nevertheless highly relevant, comprise of misinformation, fraud and coercion, the commodification of stolen data whereby the rich get richer, discriminatory facial recognition, and… uh… killer robots. No, seriously, killer robots! Needless to say, a lot of these points overlap with artists’ very real concern about their intellectual property being scraped without compensation or even consent.

Remember those Daleks from Dr. Who?! GIF via Tenor.

The third threat – again, we can look back at the Industrial Revolution as a historical precedent – is the uncompromising exploitation of labour, due to skills being demoted to monotonous and automated tasks, and craft to, well… mostly just crap (see above: threat #1). And let’s not make this all about artists: this type of hyper-exploitation is already happening in tech and textiles, just to name a few industries, and one must look no further than the Rana Plaza disaster to understand the scale of this devastation.

Moreover, not only can AI technology be the precursor to labour exploitation, it’s already the outcome: on Labour Day this year, 150 workers in Nairobi gathered to unionise, owing to extremely low wages under the traumatising workload of content moderation, which “underpin[s] the AI systems of Facebook, TikTok, ChatGPT.”

There is no turning back from the discovery of progress, especially not when considering the beneficial advancements it can bring in the field of medicine, for instance. Perhaps, in post-growth economic organisation of labour, reliant on subsistence farming and voluntary contribution, we can simply enlist robots to do the unpleasant stuff for us, while we write our little poems and paint our little pictures in our countryside cottages.

No turning back from the cave… TikTok originally via @punkreginageorge, meme via 99memes.com

If I may return for a second to Star Trek – curiously, but not coincidentally, the episode I mentioned at the beginning of this essay deals not only with Data’s sentience, but also with our responsibility as humans not to enslave other lifeforms.

We may not have conscious AI models yet but I believe there may be an incentive for developers to get there; without a certain level of self-awareness, their models can never fully replace our unique brand of human craftsmanship, unconventional problem solving, or reading the nuance in any given situation. In order to reach that level of conscious adaptability, the AI models would have to reach some amount of sentience, I’d argue, and if it is wrong to enslave and subjugate other sentient beings to hard and unwanted labour (which is very much still a problem), then it would be wrong to do the same to (more or less) sentient robots. That is not a world I want to be part of, if I’m even alive for that. In my ideal scenario, the robots and I are striking for better wages and working conditions, baby!

Human, robot, or animal… It is wrong to employ exploitative practices in order to make convenience, progress, or heightened quality of life possible, and to continue doing that, full throttle, with no ambition for reduction, once we already know that it’s wrong. It’s wrong to choose exploitative labour over people that earnestly care about what they do, and are willing to do it in order to remain within our social fabric. These are creatives and other skilled workers who cannot look into the future and know for sure that they can adapt their life’s trade and still thrive.

As for myself: this essay may seem to build upon a conclusion in which I… beg? No: take a gander at my dignity without taking it, thank you. For days, I worked on this blog post, which was not asked of me, and which I do not expect a great number of people reading. And yet, to piece it all together, I chased myself into an abyss of restless thoughts and never-ending tangents, seemingly drowning under the weight of it all – still, my heart rages at me to dig deeper. Why?

How many drag references can I possibly fit into one blog post? Dragrace Thailand Season 2, episode 9, via this reddit post.

I will continue to craft, to write, to draw, to sing, in defiance of it all, whether you choose to support me or not, and to adapt with or in spite of AI. And you can be damned sure that I will do this all while continuing to advocate for those who wish to act on the fire and hope in their hearts, but are crushed by this current system of exploitation. I hope that we can all stare at the sun together, daring it to sink before we do.

Sources:

Lioba Brückner, “How AI Solved my BIGGEST Art Problem!”, Youtube

Cade Metz, ‘The Godfather of A.I.’ Leaves Google and Warns of Danger Ahead, The New York Times

Krzysztof Pelc, AI Will Make Human Art More Valuable, Wired

Richard Connie, What the Luddites Really Fought Against, Smithsonian Magazine

Philip Prichard Henderson, William Morris, Britannica

William Morris, Useful Work versus Useless Toil, Hammersmith Socialist Society

Lindsay Varty, Why resist the sunset?, sunsetsurvivors.com

Ulrich Kirk, Martin Skov, Oliver Hulme, Mark S. Christensen, Samir Zeki, Modulation of aesthetic value by semantic context: An fMRI study, the University College London and the University of Copenhagen, sciencedirect.com

Andreas Matthias, Immanuel Kant on Means and Ends, Daily Philosophy

Mark Memmott, Shepard Fairey And AP Settle Copyright Dispute Over 'Hope' Poster, NPR News

AICAN® – “Artificial Intelligence Artist”

Garson Ormiston, Camelia Sadeghzadeh, Paul Harris, Can artificial intelligence ever be sentient?, BBC

Molly Crabapple, Restrict AI Illustration from Publishing: An Open Letter, Center for Artistic Inquiry and Reporting

Karen Hao, Training a single AI model can emit as much carbon as five cars in their lifetimes, MIT Technology Review

Chris Stokel-Walker, We Spoke To The Guy Who Created The Viral AI Image Of The Pope That Fooled The World, Buzzfeed News

Karen Hao, A college kid’s fake, AI-generated blog fooled tens of thousands. This is how he made it., MIT Technology Review

Karen Hao, We read the paper that forced Timnit Gebru out of Google. Here’s what it says., MIT Technology Review

'Godfather of AI' discusses dangers the developing technologies pose to society, PBS News Hour with Geoff Bennett, Geoffrey Hinton, Youtube

James Vincent, AI art tools Stable Diffusion and Midjourney targeted with copyright lawsuit, The Verge

Sarah Butler and Thaslima Begum, Abuses ‘still rife’: 10 years on from Bangladesh’s Rana Plaza disaster, The Guardian