Artificial, human and angelic intelligence

“The artificial in artificial intelligence is very real.”

(John Lennox, Emeritus Professor of Pure Mathematics, University of Oxford)

The difference between fantasy and imagination is that the latter engages playfully with what is possible, while the former pretends that what is fake is true. Fakery is the worst version of deceit, the devil’s opioid of choice — worse even than lying.

In recent years Satan has been distributing this narcotic through the medium of artificial intelligence. The conceit that humans can create self-conscious robots is fantasy, the pressure to believe otherwise is damaging to the (true) intuition that there is something unique about human thought and consciousness.

The global ascendancy of AI, the embrace – on grounds of conferred convenience — of the new technologies which implement it, and the transhumanist mindset driving it, pose an existential threat to all of us. Collectively these things tempt us into believing that a simulation of an experience can be as real as the genuine thing, but this is a very bad path to go down.

The sacralising of AI is pernicious, not because robots might become “more intelligent” than humans, but because they can’t. The real danger posed by the Silicon Valley evangelists is that they encourage us to think that we also are just machines – biologically and chemically advanced ones, but machines all the same. The creation of self-consciousness is their Holy Grail, and the failure to find it is taken to demand not a re-evaluation of their mission, but a re-examination of what we take human thought to be.

They never will find it, but the pressures of tenure and grant allocation entail that the search will go on. If machine consciousness is made possible only by turning humans into machines, then that’s just too bad.

I wrote my doctoral thesis on artificial intelligence and its assumptions about the metaphysics of consciousness some thirty years ago. At that time the developing systems were, it was claimed, more faithful to the neural architecture of the human brain than those they superseded. My reaction to this was: “So what?” My view is still the same, and yes, I’ve kept up with the reading. You can call a robot “conscious” if you want, but it’ll never know that it is. In fact, it’ll never even know it’s a robot.

Generative AI is impressively complex, but the complexity is algorithmic, mathematical and physical. From the point of view of what the late Sir Peter Strawson called “revisionary metaphysics”, an AI system is no more interesting than an abacus (which might, I guess, be taken to imply that an abacus is very interesting indeed). It relies on recursion and high-speed mass replication, and its relation to genuine consciousness is analogous to that between kitsch and original art.

The philosopher John Searle cut to the heart of the matter more than forty years ago. Even the most impressive AI systems are essentially collections of real-time formal operations on symbols, and these symbols will always require interpretation. Searle puts the point in the following way: syntax can never give rise to semantics.

An analogy might be helpful. Like millions of Catholics, I pray the Rosary. This can be thought of as a formal system, rules-based, and centred around the repetition of tokens of syntax (prayers). But the “Mysteries” of the Rosary – the Sorrowful, Joyful, Glorious and Luminous Mysteries — are not generated by the system. They are brought to it by myself as I enter the formal meditative space.

Of course, nothing is produced in the praying of the Rosary: it is in that sense useless, and this is what gives it meaning. It is, like philosophy, a form of contemplation whose value is self-contained, and which leaves the world intact, at least visibly.

There is a specificity to human thought which usually falls out of consideration when we think of ourselves and others in exclusively functional ways. It’s possible to make robots which behave pretty much as we do, but this is not the same thing as building something which knows what it’s doing, because what is the it that we are talking about? If there is a distinction in kind between a purely biological process and consciousness, then there is a similar distinction in kind between being conscious and being aware that it’s me that is conscious.

It is this magical specificity, the particularity of individual self-consciousness, that the AI theorists cannot account for. Human thought is not like an algorithm, tasked to produce observable behaviour. What’s important about it is the other intangible and immeasurable things that it makes possible: the ability to love and be jealous in love, to feel shame, pride, disapproval, envy, and to willingly praise gods, both real and false. Perhaps most important of all, the capacity to “get” a joke, and not simply to tell one.

The importance of specificity is explicit in natural theology. In his discussion of the angels, St. Thomas Aquinas argued that every angel must be sui generis – form its own species — owing to their proximity to God in the Order of Being. Humankind’s own specificity – what sets us apart from plant and machine — is found in this ability to reflect on ourselves.

Since the Enlightenment, an aggressive reductionism has been in play across culture and science, a project to demystify and repudiate perplexity. This is fine, but as ever there is the danger of overreach. The insistence that one thing is “nothing but” another thing runs the risk of replacing what is beautiful yet mysterious with what is dreary and unintelligible. You begin by claiming that the mind is just the brain; you end by seeing Van Gogh’s self-portrait as just a congregation of colour pigments thrown onto a canvas.

Angels are dispiritingly out of vogue, especially in the Christian tradition where Aquinas’s beautiful speculations are tolerated at best. So, consider instead the following thought experiment: a Silicon seminar, attended by androids, formed to consider the question of whether humans can really think, or if it just seems, to the robot mind, as if they can.

Absurd? What does that tell you?

A Message from TheArticle

We are the only publication that’s committed to covering every angle. We have an important contribution to make, one that’s needed now more than ever, and we need your help to continue publishing throughout these hard economic times. So please, make a donation.