Theologians of the Future

Theology is the science of the future. We will need it to understand how to enter into right relationships with things we only have an intuitive grasp of and that exceed us. You can define theology as the study of God, but that description rather begs the question. A phenomenological description might instead call it something like the body of knowledge and practice by which human beings understand right orientation toward higher intelligences and powers.

Now that intelligent machines can simulate and surpass many of the cognitive abilities of humans, the language we use to understand both the technology and the conceptual framework around AI increasingly overlaps with how we think of our own cognitive operations. The technology itself becomes too complex to describe at the level of ordinary conversation, especially outside academic contexts, so we revert to analogical terms about what AI does rather than properly descriptive terms of how it does it.

Thus, our language precludes a clear understanding of AI. At best, we vaguely sketch operative terms that serve the application of a technology handed to us; we do not seem to interact with the deepest levels of AI technology and its concepts. The reliance on analogy is not incidental. It is a symptom of the conceptual and vocabulary limits of materialist reductionism. This is due in large part to the abandonment of a long tradition of inquiry into intelligence, both divine and human, in the field of theology. At some point in the modern age we began thinking of our minds as computers. But this notion is a departure from centuries of intellectual tradition that would have raised a furrowed brow at such a reductionist metaphor. Meanwhile, philosophy and theology programs diminished in influence and status as the best and brightest funneled into STEM.

Why does this borrowing happen, and why does it happen in one direction only from mind to machine, rarely the reverse? We do not say the mind "runs inference" or "updates its weights" in ordinary speech, though the metaphor may run just as easily the other way. One answer is practical: the vocabulary of cognition is ready at hand while the vocabulary of matrix operations is recent, technical, and confined to practitioners. On this account the borrowing is an accident of convenience, and it would dissolve as the technical vocabulary becomes common, the way "bandwidth" and "algorithm" have already leaked into ordinary speech without much confusion. But this does not explain why the borrowed words are so consistently the ones that describe interiority such as understanding, attention, memory, hallucination - rather than the purely functional ones a convenience-driven borrowing would predict, like "processes" or "computes" or "outputs." We seem to reach past the neutral term for the psychological one even when the neutral term is shorter and closer at hand.

The terms are not false, exactly. They are imperfect analogies to convey what we cannot literally describe because that description is inaccessible to anyone outside a small circle of practitioners, and often insufficient even for them. The analogies help to a point, but that help conceals assumptions. To say a machine "understands" imports centuries of contested meaning about what understanding is and requires. Without a precise account of intelligence, such language obscures as much as it reveals.

The scholastics developed a vocabulary precisely for analyzing acts of understanding. For example, they distinguished the intellectus agens, the agent intellect that abstracts an intelligible form from the particulars given in sense experience, from the intellect's subsequent engagement with that form either discursively, moving step by step through inference (ratio), or immediately, grasping a truth at a glance (intellectus). These are not synonyms for "thinking", but explain distinct operations of the soul, each with its own conditions of possibility. Even the term "soul" carries a rigorously developed conceptual history, though it is now often dismissed or left undefined.

Beyond the realm of interiority, the philosophers of theology also had much to teach on ontology. They handed down a well-constructed picture of reality as a great hierarchy of beings, in which man knew his place and how to navigate the other beings according to their nature and position within it. This sort of taxonomy is wholly missing from the technological realm. Part of why we feel angst in the presence of machine intelligence is precisely because we haven't settled on what kind of thing it is.

Mapping and understanding both human intelligence and intelligence as such in any but the strictest materialistic terms may seem dubious to the faithful scientific positivists of our time, but the reach for psychological language to describe machine operations shows the limits of that framework. As artificial intelligence advances, the gap between operational knowledge and conceptual understanding becomes more pronounced.

People increasingly want to know what AI really does, what it is, and how they are or are not like that. For that, the theologians of the future will provide the vocabulary and conceptual models that technologists cannot, limited as they are by reductionist frameworks that cannot account for categorically different phenomena like mind and its operations.