Human reality remains: AI is just another tool

By Dr. Jeff Mirus ( bio - articles - email ) | Jul 10, 2026

I am reminded of the old (originally French) proverb: The more things change, the more they remain the same.* This could be applied to the constant arguments over artificial intelligence. Long-time readers will be aware that we’ve provided some fairly expert coverage of AI on CatholicCulture.org, mostly through the efforts of Dr. Thomas B. Fowler, for which I have previously provided an overview with links to the actual articles and to his comprehensive book published in 2025. Tom is currently working on an updated assessment for a prominent American magazine. But the major change in the landscape at this point seems to be that companies which invested heavily in AI development are, well, beginning to fold.

The costs of development are astonishingly high. AI is immensely resource-hungry for both its “learning” and its marginal improvements. Even the entire content of the public Internet is increasingly falling short for the adequate “training” of AI programs. And yet AI can competently perform only a few of the tasks many people expect of it. It simply cannot deliver on tasks that actually require genuine human intelligence, because even after “training” on all the resources available, artificial intelligence is missing the essential component of true intelligence. AI has no spiritual mind (i.e., no intellect, which is a spiritual property), and so it cannot really think.

In the absence of thought, however, the various AI systems usually correlate information quite well, and can be “trained” (programmed) to respond in different ways to different conditions. But they also sometimes process that information in a way that, in effect, “makes things up”. Since AI has no true cognitive abilities, it lacks, among other things, what we call “judgment”. So sometimes it processes material to manufacture statements essentially “out of the blue”. Such bizarre behavior demonstrates AI’s complete lack of a healthy personal wariness about subverting its own particular expertise. Unlike humans, AI cannot understand when it is processing information incorrectly. Various programmatic checks can be built in and tweaked, but AI does not have a “self”, so it has no self-reflection.

The spiritual mind

AI can process data only in an essentially inhuman way. For a general AI platform, this is reminiscent of students cramming for tests, using every handy resource they can find to increase their likelihood of possessing the stray bit of information they need to respond successfully to a question. They hope to be able to correlate some of their acquired data with whatever question they are asked. The difference in this example is that the student will actually learn personally from all the last-minute cramming, even though it is not the best way to acquire knowledge or commit it to long-term memory in order to foster genuine understanding.

But student cramming is actually far more efficient than “artificial intelligence”, which is aptly named by the adjective “artificial”. AI programs and machines do not possess intelligence and cannot create it, for intelligence is a spiritual property. It cannot be manufactured and installed in a machine. By contrast, the lazy student begins with intelligence by virtue of his human nature—which, after all, already contains within it the essentially spiritual qualities of intellect and will.

Note that artificial intelligence is essentially a sophisticated search and retrieval system which responds selectively to different stimuli. As such, it bypasses any form of “understanding” altogether—which is precisely why its “training” is so remarkably inefficient. Of course, this can be enormously useful in many low-level organizational systems, where genuine understanding is not required, especially when an AI machine (for this is what AI is, a process executed by a machine) can either dispense basic information to humans in order to answer questions or issue instructions to other machines in order to accomplish measurable tasks.

Now, when a reasonable answer is presented to us, we can usually recognize it, and we can test it. But AI will never score well on problems which require human judgment, except by being able to dish up very quickly what human persons have already perceived and expressed or documented about the problem in question, usually through a comprehensive exploration and mining of available sources of data. Obviously, as a research tool AI can be a significant time-saver, just as it can excel at a great many repetitive tasks that do not require good judgment or superior analytical skill or the ability (quite literally, in this case) to think outside the box.

Of course, AI systems can be “trained” (programmed) to manage and control complex machine-based processes. And in the form most of us are familiar with, AI is also good at looking up, organizing, and providing basic information or instructions for previously documented tasks. But with no genuinely intelligent way to check its results, AI systems sometimes spit out a piece of entirely bogus information. In the trade, apparently, this is called “a hallucination”.

Progress or regress?

Unfortunately, some who are interested in artificial intelligence tend to get AI mixed up with weird theories of transhumanism (which are totally bogus). More commonly, however, the challenge is simply to discern prudently where AI machinery can be made to work well, providing a speed and a precision and even, once developed, an economy in the performance of some tasks that human persons cannot match. Thus the primary challenges involved in any genuinely shrewd and well-conceived deployment of AI are simply these:

• Efficiently matching AI to the kinds of problems it can solve;
• Helping persons who are displaced by AI to transition to other kinds of work;
• Avoiding the dystopian interactions and side-effects caused by the use of AI in situations which can only frustrate, infantilize or behaviorally insult the real persons who interact with it;
• Avoiding the very basic and very damaging fallacies that “people are essentially machines” or “machines can be developed into people”, both of which are simply ludicrous.

Clearly, that last problem is potentially the most damaging, for it involves an even more rapid dissolution of the human person’s understanding of his own nature and destiny, a dissolution which tragically separates the human person from God. Over the next decade or so, I suppose, many companies will try developing and deploying AI in a variety of ways that will prove to work exceedingly poorly or, as is already beginning, will cause a backlash of human frustration. If well-grounded, this can actually be a very healthy response, especially in a society which has largely lost its understanding of human transcendence—that quality which, at root, is what makes humans essentially different from machines, and non-replicable. Only fools make the mistake of blurring that distinction but, sadly, there will be plenty of fools involved along the way.

It should go without saying that it is AI’s potential contribution to precisely this sort of foolishness which constitutes its greatest threat. Foolishness, while not without pain to ourselves and others, will continue to be, as it has always been, a recipe for failure; and AI-prompted foolishness about the true nature of man is just another form of that diabolic foolishness that tends to separate us from our Creator.

On the other hand, if the smoke of Satan does not become completely blinding, the AI revolution will leave humanity no worse off than it already is, and it may provide significant human benefits along with whatever challenges may arise. Even if it is as disruptive to human habits and patterns of life as the industrial revolution—which I doubt—we can weather its social and economic disruption, and offer assistance to those who are adversely affected. Certainly a great many of those who are spearheading the AI revolution are going to lose their shirts, but by God’s grace they may still save their souls—if only they are open to the grace they need to thoroughly understand the difference between the man and the machine.

For that is the first step to enhancing true human intelligence. A beneficial progress for humanity requires not only material improvements but spiritual growth leading to a better discernment of the ends and means appropriate to the conditions and possibilities in each time and place. In its true human context, artificial intelligence is just another tool that ought to be used more wisely than not. But what we have learned about AI so far is that, even with all the resources in the world available, it cannot substitute for human intelligence. As I mentioned: The more things change, the more they remain the same.

---

* Plus ça change, plus c’est la même chose.