Weirdly I tend to forget what motivates my views on AI sometimes. I’m ultra critical, obviously, sometimes outright abolitionist, but as I spend so much time focused on it even my own driving motives tend to become obscured beneath papers and articles and research. I can point to the bad, the critiques, the failings and the harms but at times it almost becomes inert. The building blocks of an argument I know can be made even without much intention around why I specifically undertook it.
Thinking about that kind of drew me back to first principles. Beyond the academic obligation what is my core motivation to oppose AI? At heart I think it’s the inherent opposition to the human that so many of the technologies of AI represent. In the arts, in decision making, in surveillance even – the driving inspiration for AI mobilisations is disdain. It’s an overarching sense that the human is insufficient, our creations, our behaviours, our existence even – they’re all inherently faulted and in need of perfection. And in a way that’s true. We are an incredibly faulted species even as far as our own declared intentions go but where we do succeed it’s out of a chaotic mix of intent, chance, empathy, exploration and ideas which we act upon and embody with often only the vaguest notion of what’s actually going on. The idea that as a species, or a culture, what we really lack is the machinic intervention to define and pull us to some definite, tangible goal is a fundamental misunderstanding of what we are.
Art is always the easiest example to explore for that. GAI says ‘here is the end point, this platform will enable you to reach it’. That’s a fundamentally anti-human logic. Even if your aspiration is to paint like Picasso then in the process of pursuing it your innate chaos will intervene. Perhaps you’ll never reach that goal, perhaps you’ll drift to something else, perhaps in your abject failure you’ll become something else entirely. The vast and intricate web of emotions, ideas and distractions that consume the human experience may lead you anywhere. The generation of a product which narrows down the definition of ‘right’ to adhere to its own limitations won’t.
Even beyond the artistic though the same logic applies. We routinely make bad decisions about welfare for example (benefits, medical care, mental health etc). The AI intervention there is to say ‘feed us your data and the right decision will be made, unhindered by the chaos of human bias and limitations*’. How often is the right decision clear though? Where positive ends are reached it’s often against all declared logic. It’s where care or support is offered despite systemic (and even personal) declarations that it’s neither needed nor deserved but the chaotic perceptions of the networked human – loaded with latent memories, understandings and empathies – decides otherwise. There is no linear, statistical, logical order to that and so the assertion that AI can offer one is less liberatory and positive than denigrating to the unspoken and misunderstood.
Art and care both rely on the contextual, temporal, spatial human experience. There is no dataset sufficient to emulate that, there is no model so perfectly balanced as to understand why person x can intuit why person y needs care despite all external indicators. There is no GAI which can explain why an atrocious drawing of a six legged dog will draw out more joy from the observer than a near infinite supply of perfectly produced photo-realistic puppies. To improve the human, insofar as that’s ever possible, there must be an understanding of what’s not understood, a recognition of the nonsensical and counter-factual, an augmentation which increases the scope of our best impulses without ever declaring authority over the end point they reach or the best path to pursue them through. Judgements which, even if routinely fallible, belong to the human above all else.
It’s that dismissal of the human that, ultimately, frames my views on AI. For all of our myriad imperfections I still oppose the technosolutions that seek to further degrade and denigrate us. For all the tools we may use to mitigate our own capacity for mistakes the ones we’ll never need are the ones which quietly declare us fundamentally incapable and our natures fundamentally faulted. Whatever ends we may pursue, rightly or wrongly, we who can so rarely define them with any certainty absolutely don’t need a machinic voice declaring what they must be before circumventing our agency to discover that for ourselves.
None of this is meant to place any thoughts of mine at the head of a hierarchy of concerns about AI. It does far more immediate harms all the time. And it’s not a particularly academic or profound point to make, I’m only talking about my own motivations which come from my own positionality and my own perceptions. Still, for me AI is an insult. It’s a judgment which says the necessary tools and requirements of humanity are beyond humanity, and without augmentation from a remote and centralized force, we are eternally insufficient. That’s a thing worth resisting.
– Dylan
*Nonsense in itself, but taking the motivations at occasionally declared face value.
Leave a Reply