Image generated by the author using OpenArt. Prompt: Detailed, HD realistic colourful representation of a machine learning algorithm in the style of a Baroque painting.

Platform Realism, AI Aesthetics, and the Metaphors We Live By

What intelligence do we actually refer to when we speak of artificial intelligence? Can a machine learn? Metaphors are heuristic tools that allow us to grasp concepts through the use of language that is familiar to us. “The concept of the metaphor is quite powerful, because it cognitively taps into different portions of our being and our neural systems, and interacts with memory and visual perception. The human species has always worked through metaphor, through the complication of visuality and ephemeral concepts.”

Historically, especially in communal and Indigenous cultures, transgenerational knowledge has been transferred through fables, parables, and poetry. These conceptual ideas have allowed individuals to see themselves within the narrative. For example, consider parables. Amaro recounts one where a man is crossing a lake with a scorpion in his pocket. This imagery helps us connect with the story on a deeper level; “Maybe that scorpion is another human being or it’s a process or it’s a political structure.”

These types of structures are also present within AI and machine learning, and they ostensibly serve the same type of function, by breaking our perceptual lens and allowing us to feel, experience, be in the moment and then therefore gain a sense of agency in which to participate in that moment itself. “And this is why I'm so against the debate in AI about what's real and what's not real,” says Amaro. “Because the scorpion in your pocket doesn't have to be real for you to learn the lesson, because the real transformation will come with your connection to a larger body of wisdom and knowledge itself.”

Sycophants often tout the supposed ability of AI to “realistically” reproduce reality[1]—as if such a process can ever be carried out. Yet these claims obscure how the emergence of such aesthetics ascertains the status quo; the process by which a written prompt is transmuted into an image is by no means neutral, but rather subscribes to algorithmic logics, which in turn attempt to, in the words of Flavia Dzodan, “render the complex and particular into legible, aspirational forms.”[3]

I stirred the subject of my conversation with Amaro toward the viral proliferation of an AI-generated image of what is ostensibly supposed to be a refugee camp, though what is absent from it is more important than what is present. Bearing the text “ALL EYES ON RAFAH”,  the image is bereft of any human element. Shared millions of times, it became memefied and subsequently drew critique for its sedimentation of a sanitised representation of genocide and dispossession. Admittedly, Amaro’s answer takes me by surprise.

“When you make a caricature of a political statement in a cartoon, that's not a real depiction either. But it's distributing a message,” he says. “We don't need AI to tell us that there are millions of people who are displaced and in danger. I don't care if it represents reality, because I know reality; I don't need to see it. Before AI, we had the propaganda poster, the placard with a funny message or something else like that might not even depict reality,” he continues, “but we don't blame the cardboard for saying, hey, you're not depicting the world as it really is. And it's not representative of the world.” For Amaro, the problem boils down to subsuming this into an issue of what's the truth, what's a lie. “While we're debating on whether this image is real or not, there are babies dying. Who cares? Who cares if this image is real or not?” 

This seems to be a leitmotif in discourse around “new” technologies: “We ignore the real issue, and then we subsume the most superficial of that issue, and then that's what we debate. It's a moment of safety because we can't face the complication of what's going on. And what Gaza is currently representing now is a particular Western crisis.

“Humanism has prided and valued itself on human rights, on the capability of expressing rights, on what it thought was life itself. Yet and now in many cases it's supporting the destruction of life once again. So instead of us addressing that, we want to look at an image and do an analysis of, you know, was that tint generated by AI or was it a real tint? Is it a distraction?”

History has already shown that what we see, we're not ready to see reality. We've seen it historically. How many more times do we have to see it? We cross homeless people and rough sleepers all the time, every day.”

“The problem is that we're demanding people have solutions when we ourselves don't even have a solution,” says Amaro. “Do you know how many times I get asked, well, what's the solution to race? If I had a solution to race, I'd be on a yacht right now.” The idea that a single human being, let alone a tech giant, ought to possess that singular panacea is absurd. “I'm not giving that solution to Mark Zuckerberg, right? I don't expect him to have it, because no single human being should possess that much power.”

Towards a More-Than-Human View of Technology

Cultural imaginaries around AI oscillate between production and destruction. Writer Jason Parham writes that “[the] awe of what AI will achieve is undeniable. But so is the fear.” When not concerned with how it will help make our lives easier, AI is shrouded in a public epistemology of fear: the technology can be weaponised against humanity, peace or democracy, it reproduces harmful norms. Writer Jason Parham writes that “[g]enerative AI pulls from—learns from—the ugliness of human error.” If it isn’t discussed as a harbinger of catastrophe, AI is understood as a solutionist panacea, a ‘surrogate’ technology, a tool on which we (ought to) rely to show us reality, do the chores, solve our problems. Even leftist outlooks on automation hold this view.

It’s of course tempting to look towards technology for solutions, but Amaro is not concerned with what it can do for us. “We're always taking this anthropocentric position, assuming we’re this apex of species on a biodiverse planet, by saying that it’s Other’s job to teach us something. I'm trying to reposition our relationship with technology away from that idea.” 

This is of course easier said than done. Powerful, heavy, financially viable conglomerates have decided what these technologies will do for humanity. How do we tip the scales back towards alternative forms of sociality, in tandem with these technologies? “What we have to do, and here's the kicker, here's the problem, is we have to get rid of the aim, we have to get rid of the idea that these technologies are useful only for commercial aims,” says Amaro. 

In spite of all the currency that has been spent, and all the talent and brain energy, the only thing that we can really produce is a sophisticated language model that can write an email for us. “Because that's sellable,” Amaro says. “Whereas you see alternative practices, small pockets of communities, First Peoples, Inuit Peoples, Indigenous communities, Black, Brown folks, trans folks, artists and makers who are trying to break that system and come up with alternative languages in which these technologies can express something different, imagine a world where the principal concern of technology isn't fattening a bank account.”

Amaro likens this with the situation of someone squatting in your house, leaving you to spend the rest of your life standing outside of your own home. “This is why I want people to not do machine learning,” he says. “I want people to learn what it is, because I want people to go into the house and see who's real and who's a fiction. To see that actually the individuals we see as all-powerful are all-powerful for certain reasons, and we need to detach ourselves from the idea that those systems are impossible to disrupt.”

Ultimately, in this crisis of imagination, we must, as author Lola Olufemi suggests[3], navigate the space between what is and what could be. We need to cultivate the ability to imagine otherwise—to envision the possibility of disrupting the systems we take for granted.