Trying to conceive of the future of AI presents a number of challenges, especially for those driven solely by interest and not profit. On the one hand, there is an overarching urge to accept and even over-hype AI’s potential. But in the opposite direction lies a mass of popular narratives and cynical postulations that the destruction of humanity will come about in the ‘Age of AI.’ The tendencies, impulses, and assumptions that fuel these discussions often belong to the people who possess the capital or clout to determine the future of AI. Which leaves us with this stunning reality: the future of AI is tangled up in all the fear and the obsession.
With this in mind, thoughtful articulations like Zylinska’s in AI Art: Machine Visions and Warped Dreams function as an essential posture towards AI as questions about intelligence, creativity, and ethics coalesce in this discourse. There are a few thoughts in particular that I want to hone in on that pertain to Zylinska’s text, but I want first to acknowledge that my areas of intense interest do not cover the entirety of AI Art: Machine Visions and Warped Dreams. There are extensive sections of the text that I am not addressing here, but that also deserve attention. If you are interested in Zylinska’s book, you can find it here.
Intelligence and creativity are commonly used terms that attract a great number of assumptions and biases. When thinking of these terms, the automatic metric for comparison is entirely human. However, if a single thread of thought is pulled at hard enough our models of intelligence and creativity are liable to unwind. Intelligence as expressed by humanity continues to evolve as we do, but human activity for the most part has always been technical and thus somewhat ‘artificially intelligent’ as Zylinska explains (13). The use of the wheel can, for example, be described as artificial intelligence because intelligence was applied in the use of an artificial object. This stretches the bar away from requiring that AI be a sentient entity to recognizing that Deep Learning is a function of mathematics and infrastructure, first and foremost.
With this groundwork laid, AI loses some of its mystique. It is certainly complex and difficult to develop, but it is not earth-shatteringly new. Paradoxically, it is both something new and something similar – another link in the long chain of humanities’ efforts to expand and adapt. The catch is that at some point or another, AI might outpace us, per say. Our limited perspectives as humans temper and form our ideas of what constitutes intelligence (13). Which means that were AI ever to produce an intelligence of its own, we are likely to overlook or ignore it entirely. As Zylinska states, “AI’s intelligence may take the form that is not only superior to that of the human, but also unrecognizable by humans as intelligence” (34).
Likewise, creativity is bizarrely linked to intelligence since it appears to stem outward from it. Creativity may in fact be as simple as a change in an organism’s response to its environment (65). With limitless existence and established intelligence, an entity could theoretically eternally expand its creative potential. But does the endlessness of this process enable or inhibit the impact of creative expression? Zylinska expresses an idea that has haunted me. She explains that the work of a GAN, for example, could become “an ouroboros-like circle of random variations” culminating in “the pointless production of difference” (72). Does ultimate variety matter if the viewer is inundated with information and consequently becomes unengaged? I hope that I can investigate this more thoroughly, but it may be something that can only be revealed with time.
In the meantime, we must address the ethical considerations surrounding AI. There is a substantial push to prepare for the repercussions of a hostile Artificial Intelligence. In fact, “Combatting climate change as such is therefore seen as less urgent by the majority of investors than combatting robots” (43). This mentality plays more to the desires and egos of investors than addressing the already overwhelming problem of climate change does. Additionally, these individuals can continue to make money off of security protocols surrounding robots and AI ad infinitum whereas confronting climate change requires the abandonment of many capitalist ideals including the concept of never-ending growth and expansion.
The conundrum is that these individuals shape the future of AI with their fears and fixations. They control the capital needed for technological development, but are hesitant to share that power. Which leaves us with a plundered world still controlled by a few, even in the theoretical Age of AI. If our future is to be drastically shaped by AI, it is telling how control of technical systems is already so centralized. We have the façade of accessibility through cloud computing and hosting services, but the haunting reality is that the supply of these systems could be ripped from the hands of the public at any moment. There is no apparent end to this in sight and labor opportunities and conditions will continue to be controlled by the few for the foreseeable future (124).
While this knowledge gnaws at the back of my mind, a seed of a thought comforts me. Storytelling and art making are more than just an escape and “may even be the first step on the way to ethical, or responsible, AI” (30). Using these methodologies, we can reimagine our future and conceive of a world beyond our perceived limitations. We can also further educate society about AI and instigate new expectations. Evolutions of expectation can influence the future, or so I hope. Zylinska’s text inspired and challenged my ideals and I hope through my participation in these endeavors I will be able to inspire a similar shift in people’s desires and dreams about AI.
Zylinska, Joanna. AI Art: Machine Visions and Warped Dreams. London: Open Humanities Press, 2020.