The Threat of Automation

I recently incited wrath in the comments section of an Instagram post.

The account was a cinematic dreamscape of AI-generated clips which rendered me (forgive the pun) in awe and unnerved. These videos, while captivating, were grotesque. Cherub-cheeked trolls towers over little girls in lederhosen, scaled 100:1, against a backdrop of a mossy fairytale forest. A midcentury housewife dressed like Donna Reed in a nightgown but with a green alien face surrounded by pastel pink everything slaps her hand over her vintage alarm clock. As I compose these word-soup sentences in an effort to relate the imagery of these videos, it occurs to me that I may be reverse-engineering the prompt used to create them in the first place.

The visual fidelity of these videos make me do a double-take to ensure they are, in fact, AI. As of late, AI image and motion graphic technology has been progressing at such a dizzying pace that many of us who try to keep up with these advancements are left gob smacked at creations featured on these mysterious AI art profiles popping up on Instagram. I had to do a double take to ensure that it was AI and not a huge studio producing them, because it had not been long before that I recall AI motion imagery giving much more, shall we say, first-season-of-The-Simpsons vibes.

I feel their strangeness leaning hard against the deliberateness of their manufacture. They are both familiar and unique, haunting and charming. They are dreamy in a way that utterly reminds me of Rene Magritte and the surrealists, who plunged into the depths of their subconscious in order to resurface unexpected salvages from the seafloor of their minds and spirits, to realize a vision of something both familiar and unknown that will move the viewer, even if only by knocking them off their axis. These AI clips threw me off my axis, and like great surrealist works they summoned something familiar in an unfamiliar context.

You’ve probably heard the buzzy expression “uncanny valley.” It popped into the cultural lexicon around the same time as AI, and is used in reference to the phenomenon whereby a computer-generated figure or humanoid robot bearing a near-identical resemblance to a human being arouses a sense of unease or revulsion in the person viewing it (Oxford). One theory about this phenomenon is that it stems from a human evolutionary trait called pathogen avoidance, wherein we are programmed to feel unease and revulsion at humanoid faces because our brains register them similarly to that of a corpse, which can carry harmful pathogens. It is our brain’s way of protecting us; telling us that something deceptive and dangerous behind that familiar face.

The term is certainly appropriate to describe these videos. In this instance, it wasn’t a human face, but the resemblances of tone and form to ubiquitous creative works. I did not need to investigate which models created these images to know that they have been fed imagery of real-life creators I admire. Undoubtedly sourced were images from the works of Jim Henson and Tim Burton, who has been an outspoken critic of how emotionally violating, invasive, and insidious is the intellectual theft this technology operates on.

The sentiments expressed in the comments were overwhelmingly negative. People’s tones ranged from skeptical to fearful to irate. Comments about the creative destitution of the person who generated these clips. The threats AI imagery technology poses to “real” artists. The overall threats of AI to humanity. Some Matrix jokes. The usual echo chamber of that typically accompanies any controversial Instagram post with an open comments section. The account holder responded to a few of these comments with a different iteration of the same sentiment; “I am an artist. I imagined these images and used a new tool to create them. If it’s not for you, don’t look at AI art.”

Returning to the wrath I incited; I believe I said something along the lines of “When photography emerged in the middle of the 19th century, the automated nature of the medium provoked very similar fears and critiques as some of the thoughts shared in this comments section, and many people decried it as not being art.” You can imagine how that was received. This comment was not an argument in defense of AI imagery. It was a simple statement of historically recorded factual information, an observation of parallels between the past and present. It was not an endorsement of AI, nor meant to encompass the complexities of the arguments for or against AI art. None of the ethical considerations relating to original artists, nor the threats to the livelihoods of creatives today are encompassed in that comment. None of the considerations about how such technology may shape creativity in generations to come.

I will say this: It is an evocative medium. As I consumed those videos; this strange robo-art, I had the range of feelings I can expect to have when experiencing art. They were consuming. They were thought-provoking. They were a sensory bomb of color, tone, and texture. I felt unnerved and confused and fearful and offput. Art has never had to be ethical to be art. It’s never had to appeal universally to be art. It’s never needed to be considered safe or good for society. Art pushes boundaries. Art challenges people’s notions of what art is. It has never historically been the role of a generation to determine for the upcoming ones what art will be for them.

Like so many aspects of technology that people hate, AI is unfamiliar. It is radical, and it will change the world. But it is completely precented for technology to incite fear, indignation, and range. Humans are programmed to desire the familiar. People believe that AI art is not human-centered, but it is as it currently exists, and our perceptions about automation in technology is relative to the time we live in. Speculation about how its potential to evolve to the point of sentience and autonomy is another topic and requires serious consideration. To that end I will say, the first humans to harvest fire would have been doing themselves a disservice to reject it because of its dangerous nature. The flame of AI can light the way forward for humanity or burn it to the ground. Understanding AI is the only way to influence those outcomes.

For as long as humans have been researching and developing, we have been creating tools that change how we create. And the shakeups to industry that come with new technology have historically been  met with hesitation. The only thing unprecedented here is how fast these advancements are happening. The tidal wave of AI will not be stopped, much as many of us would like it to. I’m not necessarily saying I am riding the wave, but I’m not going to bury my head in the sand either. It’s important that we have these conversations about AI ethics, that we regulate how these models are developed. Because I know one thing for sure. Surrealists may reach into the depths of their own souls to create their works, but I don’t want to see a robot doing the same to Tim Burton.