I’ve decided to start a blog here to discuss some of my thoughts on current events and how one might apply Heathen philosophy and spirituality when considering such matters. These blog posts will eventually be gathered into longer articles; it’s very difficult for me to write short pieces because I keep thinking more about them! But here I’m beginning my attempt to address large topics in small chunks, starting with this series of posts I’m calling “Soul in a Box.” It may not seem that I am bringing Heathen thought to bear here, since much of what I’m saying is also being said by others who are not Heathen (as well as by other Heathens). Please be patient, though: we’ll get there!
(Note: After writing this post and giving this series the title “Soul in a Box,” I discovered there is something else by that name, relating to a computer game. I’m still keeping this title for my series of posts here; you will see why as you read, and may come to agree with the irony I see in the use of that title for an electronic game.)
The ongoing debate about whether AI / LLMs and their ilk are equivalent to humans illustrates, to my mind, the severe limitations of our culture’s general idea of what it is to be ‘human.’ This limited perception, in turn, channels the broad, multidimensional expanse of human-ness itself into a box with defined and limited parameters. Then the purveyors of so-called human machines can point to the similarities between the capacities of those machines and the capacities of humans-living-in-boxes and proclaim: “See: our machines can do everything that humans can!”
And in the case of the so-called superintelligent AI systems—which are supposedly hovering on the horizon—the idea is that those will be so much ‘better’ than humans that they can be considered ‘gods’. (Excuse me while I recover from laughing, weeping, and screaming at the same time…something which AI is unlikely to do!) This is what results from a couple of generations of people who have grown up creating and playing video and computer games—this desperate, and dangerous, perception that we humans are all programmed to behave like machines contained within boxes, and that these boxes, furthermore, are equivalent to the ‘real world.’
The effort to equate machines with humans doesn’t tell us how great the machines are; it tells us how limited the modern understanding of human-ness is, in the minds of those who propose such an equality. And even more, it tells us how limited their understanding of life itself is—life as all beings experience it: humans, non-human earthly beings, and spiritual beings as well.
I am not denying the power of machine information processing and ratiocination; it is indeed amazing and has many applications for today’s world, both helpful applications and horrifying ones. Emergent properties arise from the machines that mimic irrational or non-rational as well as rational human behavior, which is not surprising since the machines are trained on material exhibiting such behaviors.
One of the subtle philosophical traps that arises in this debate about whether AI systems are equivalent to humans is the distinction between ‘what we do’ and ‘what we are’. Those who argue that machines are as good as, or better than, humans are doing so on the basis of what the machines can do. They are subtly implying that there is therefore an equivalence between ‘Doing’ and ‘Being,’ which then makes the machines the same as, or better than, humans in some ontological sense—in the nature of our respective existences.
This is a mechanistic view of living beings as a whole: in this view, we are what we do, and that’s all there is to it. If someone or something does things better than we do, that supposedly makes them better in an existential sense. Modern humans extend this kind of value judgement to people (‘workers, performers’) and also to Nature and the beings and phenomena of Nature, which are either useful (to us) or nuisances or dangers (to us). We become very frustrated when people, beings and processes of Nature behave in ways we don’t want, while valuing them in a transactional way when they do behave as we want.
It’s no wonder that we have become so focused on machines ever since the beginning of the Industrial Age: in machines, we’ve created things that reliably do what we want as long as they are functioning as designed, unlike living beings and natural processes which go their own ways. The irony is that now we are verging on becoming the machines of the machines.
The way things are going in work spaces, public spaces, and private spaces, AI / LLMs and their capacities such as surveillance and response are subtly, and not so subtly, pushing us to adapt our behavior to what the machines require of us. The machines of the Industrial Age forced workers to adjust to using machines in a physical way while at work, instead of doing the work at their own pace and in their own way. AI / LLMs, smart systems, etc, are pushing us to adjust our entire way of thinking and our behaviors in all domains of life in order to fit in with their dominance of our world.
There are many obvious and daunting examples of this, but here I want to give one that may seem small and silly. Many people have AI assistants or robo-assistants or ‘smart-homes,’ etc., that will do all kinds of simple (and complicated) things for them: turn lights, heating or cooling, or ovens on or off, for example. This is done by the person’s voice giving a command. The person believes they are giving a command to a machine that will do what they want.
There’s another way to look at this, though. Think of it like a person training a dog to bark or roll over on command. The person won’t give the dog a treat until the dog does what the person wants it to. It’s the same with smart systems running someone’s home or car. The system ‘rewards’ the person by doing as they request, like turning on lights or music. But often, that’s not all that’s happening. The system may also be collecting data about the person and using that data to exploit them through advertising and scams, or surveilling them for other purposes. The smart-system is thus training people to ‘bark’ (speak an instruction) and then they get a ‘biscuit’ (the light or the music turns on). This exploitation is becoming ever more prevalent and sophisticated, and it is training people to give them the data and control that the system wants.
Right now, the machine-system is ultimately controlled by, and benefiting, other people, businesses, and governmental bodies, but it’s not inconceivable that ultimately the machine-system itself will be running the show. People give up their independence this way: instead of simply walking over and switching on the light, they submit to ‘being trained’ by their smart systems to become dependent on them. Simply getting in one’s car and driving, these days, plugs one into a massive data collection and surveillance system, and the more ‘smart systems’ are involved, the worse it gets.
It’s bad enough when we simply can’t avoid it, but it gets worse when we actively choose to use devices and systems and conveniences that take over our thinking, our actions, our reactions, our independence from us. We think we are controlling our environment this way, but really the systems are subtly training us to behave in a way that benefits them, like a person training a dog with rewards.
The claim is that AIs / LLMs are, or soon will be, somehow duplicating and then surpassing the essence of what it is to be human. In order to make such a claim, one must have an idea of ‘what it is to be human,’ and in fact, ‘what it is to be a god,’ too, to address the idea of ‘AI gods.’ This is what I want to explore in this series of posts that I’m calling ‘Soul in a Box.’ I’ll be back soon with the next post.