Pretense and Confusion

There is a very human tendency to deify things, especially the things we ourselves have made.

Supposing one were a thing after all – a thing designed and invented by Someone Else and valued for qualities quite different from what one had decided to regard as one’s own true self?
– C.S. Lewis, “That Hideous Strength”

In Russel Kirk’s posthumous memoir, The Sword of Imagination, he shares a remark made by one of his friends, who had described the study of psychology as “the deification of the ulterior motive”.  I laughed when I read that, because it contains an element of truth and, as someone once said, “only the truth is funny”.

There is a very human tendency to deify things, especially the things of our own devising. This dark inclination was warned against repeatedly by the biblical writers, from Moses to the apostle Paul. They continually waved the red flag over our propensity to make gods out of our own innovations.

In a recent and fascinating post on Substack, writer N.S. Lyons suggested that, for all the talk about the prophetic insights of George Orwell and Aldous Huxley (1984 and Brave New World, respectively), it was C.S. Lewis, in That Hideous Strength, who offered the most prescient hints about our current moment. In that book, Lewis imagines a world in which the cultural elite have devoted themselves to “the Head” which, as the story unfolds, turns out to literally be a decapitated human head. Enamored with their own technological prowess, the elites maintain “the Head”, mounted by a bracket on the wall in a secret room. Ostensibly kept “alive” by an amalgamation of electricity, pumps and vacuum tubes, it can be made to utter words by pumping artificial saliva into its mouth, whereupon it will wheeze out instructions for its followers, even as it drools onto the floor. Lewis, I think, paints this ghastly picture as a kind of mockery of the prideful pretensions that often attend human ingenuity. Later, of course, we discover that the words wheezed out by the talking head actually have a demonic origin, with the head mounted on the wall serving as a vehicle for misdirecting the attention of the elite devotees, not unlike the visual misdirection usually used by magicians.

All of this was running through my mind as I listened to Yuval Harari’s most recent prognostications about artificial intelligence. Harari is perhaps the closest thing intellectuals have to a pop star.  His books have sold millions, and he has the ear of many of the most politically and culturally connected people in the world. How he thinks about things will eventually become the way many other people think about things.

The way an argument is framed is often more decisive than the argument itself. And that is certainly true in Harari’s case. He begins his talk by suggesting that AI possesses actual agency. He describes it as potential first alien life form, or inorganic agent, in world history. While Harari admits that current AI lacks consciousness or feelings, he nevertheless proceeds to speak of AI as if it has its own intentions, apart from its human inventors:

The new AI tools are gaining the ability to develop deep and intimate relationships with human beings. (5:24)…The most important aspect of the current phase of the on-going AI revolution is that AI is gaining mastery of language. (6:09)…By gaining mastery of language, AI is seizing the master key, unlocking the doors of our institutions, from banks to temples. (6:20)…AI has just hacked the operating system of human civilization. (6:45)

Gaining. Seizing. Hacking. AI, Harari implies, is doing these things itself - of its own accord. Harari, like a growing number of intellectuals, is clearly in awe of AI’s ability to wheeze out words. And the linguistic capabilities of the language models that have emerged are impressive, make no mistake. But they are not active agents in pursuit of their own intentions, notwithstanding Harari’s narrative framing. Alas, the reach of his presentation skills apparently exceeds his grasp of technology.

The question of intelligent agency quietly inhabits much of our cultural discussion surrounding AI. The embedded but unspoken premise is that human beings are mere passive observers of something proceeding according to its own intent. There is a curious emphasis on what AI might do, with little discussion about what human beings themselves are doing. The subtle framing of the discussion in this way, suggesting AI has its own intentions and inevitability, is eerily analogous to the way Lewis’ elite devotees conceived of their own beloved “Head”.

We are the ones building the data centers, writing the code, designing the semi-conductors. We are the ones provisioning the hundreds of megawatts of power needed just to analyze and emit paragraphs of readable text. Yet we are also the ones being subtly nudged toward the belief that it is somehow AI which is, on its own, doing all of this gaining and seizing and hacking.

The popular conception of AI is on a trajectory to make it the golden calf of the 21st century. Like the Israelites of old, in our crazed confusion over what we are even doing here, many seem determined to actively choose self-deception over any possibility that what we are, and what we are for, was ordained long before our arrival. The media is actively unhelpful in this regard, with its consistent narrative framing in which AI is described in awestruck, mystical tones. We would be better served to take responsibility for our own actions, and to stop pretending that we ourselves are not the ones behind the AI curtain. AI isn’t the one with agency here -- we are.

---

This post first appeared in Salvo