President Trump wants to rename Artificial Intelligence. Good. So do I.
On September 19, Trump called the name “Artificial Intelligence” inaccurate and ineloquent and proposed alternatives: Superior Intelligence, Extreme Intelligence, or Supreme Intelligence. He put the question to a public poll. The following day, Supreme was dropped and the choice narrowed to Superior or Extreme.¹ His willingness to question the received name raises a more fundamental question: What exactly are we naming?
It is a question I have been asking for several months. On June 10, I opened my Naming the Machine series with a proposition older than computers: we encounter reality first as gift, then seek to understand it, then name it, and finally learn to steward it. Adam names the creatures in Genesis. Names are not merely stickers. A name tells us something about what a thing is, and what we believe a thing to be affects how we treat it.²
That is why I welcome Trump’s willingness to question the received name while remaining unconvinced by the alternatives. Superior Intelligence. Extreme Intelligence. Supreme Intelligence. Each changes the adjective. My concern is with the noun.
It is still not intelligence.
The word artificial points us toward artificium, the craftsman’s art, and so toward what that art produces: something made. The machine is a made thing, the product of human ingenuity applied to hardware, algorithms, data, instructions, and enormous computational power. It can identify patterns, make predictions, generate language and images, compare information, and produce results that surprise even the people using it.
None of that makes artifact a term of disparagement. A cathedral is an artifact. So is a violin, a telescope, and a particle accelerator. Human beings make astonishing things.
Capability can be extraordinary without changing the category.
The harder word is intelligence. In the older philosophical tradition behind the word, intelligere points toward understanding, toward an intellect that apprehends and judges.³ That is considerably more than producing an output that looks intelligent. Before asking whether the machine possesses superior, extreme, or supreme intelligence, we should ask the prior question: Is intelligence what we are looking at?
There are several ways of speaking about the thing without requiring one word to do every job. Artifact tells us what kind of thing it is: something made. Model is the language commonly used for the computational systems themselves. Tool tells us something about the machine’s relationship to the human person. Taken together, they give us a tool for reasoning: a computational artifact made by people to assist them in working with language, information, patterns, comparison, and inference.
A tool keeps the human being in the sentence.
That distinction matters. In July, in “Stewarding the Machine,” I considered the remarkable case of an OpenAI model producing new mathematical configurations in work on the Erdős unit distance problem. Tim Gowers called the result “a milestone in AI mathematics.”⁴ But human mathematicians still had to evaluate the work, understand its significance, digest the proof, and place it within the larger field. The machine contributed something valuable. It did not assume responsibility for understanding what it had produced.
That responsibility remained with human beings.
The same principle follows the machine out of the computer and into the physical world. The made thing does not hover somewhere in the ether. The systems we casually call AI require data centers, and data centers occupy land. They require electricity, water, cooling systems, infrastructure, capital, and eventually decommissioning.
The tool is not the plant.
That distinction has become increasingly important in Texas. Since June, Governor Greg Abbott has directed state agencies to address several consequences of data-center development. On June 10, he directed the Public Utility Commission of Texas and ERCOT to take steps intended to prevent the infrastructure costs associated with data centers from being passed on to residential ratepayers. On August 3, he ordered a comprehensive audit of every data-center project advancing through ERCOT’s interconnection process before any project could move forward. That audit requires information about public financial assistance, electricity demand and generation, water sources and consumption, cooling technology, effects on neighboring communities, and ownership and control. On September 14, Abbott directed the Texas Water Development Board to enforce existing water-use reporting requirements for data centers.⁵
These are the kinds of questions stewardship requires. Supporting the development of a powerful technology does not require surrendering judgment about every physical project built to support it.
Welcoming the tool and judging the plant can be the same civic act.
Which brings us back to the name.
A made thing can be astonishing. It can extend human capacity. It can surprise us. It can help us discover something we did not know before. None of those things requires us to mistake the artifact for the person using it, or to surrender the human obligations of understanding, judgment, and responsibility.
Technology does not remove the obligation to judge. It increases it.
Name the machine accurately. Learn the tool. Judge the plant.
Footnotes
1. Donald J. Trump (@realDonaldTrump), Truth Social posts, September 19–20, 2026. On September 19, Trump described “Artificial Intelligence” as “inaccurate, and very ineloquent,” proposed “Superior Intelligence,” “Extreme Intelligence,” and “Supreme Intelligence,” and put the three alternatives to a public vote. On September 20, he removed “Supreme Intelligence” and opened a new vote between “Superior Intelligence” and “Extreme Intelligence.” See also reporting and archived copies of the September 19–20 posts.
2. Karen Early, “Naming the Machine: Artificial Intelligence and the Human Person,” Fides et Ratio, June 10, 2026; Gen. 2:19–20 (RSVCE).
3. Thomas Aquinas, Summa Theologiae, I, q. 79, a. 8; II-II, q. 8, a. 1. Aquinas writes that “to understand is simply to apprehend intelligible truth” and describes human reasoning as proceeding from things understood and returning “by way of judgment” to first principles. In II-II, q. 8, a. 1, he explains intelligere as intus legere, “to read inwardly.” The latter is Aquinas’s own etymological explanation and is not presented here as the modern philological derivation of the Latin word.
4. Karen Early, “Stewarding the Machine: Prudence, Subsidiarity, and Living Well with Artificial Intelligence,” Fides et Ratio, July 1, 2026; OpenAI, “An OpenAI Model Has Disproved a Central Conjecture in Discrete Geometry,” May 20, 2026; Noga Alon, Thomas F. Bloom, W. T. Gowers, Daniel Litt, Will Sawin, Arul Shankar, Jacob Tsimerman, Victor Wang, and Melanie Matchett Wood, “Remarks on the Disproof of the Unit Distance Conjecture,” arXiv:2605.20695 [math.CO], May 20, 2026, esp. secs. 3–5. The companion paper describes itself as presenting a “digested, human-verified version” of the model-generated counterexample. Gowers describes the result as “a milestone in AI mathematics.”
5. Office of the Governor of Texas, “Governor Abbott Directs PUC And ERCOT To Shield Texans From Data Center Infrastructure Costs,” press release, June 10, 2026; Office of the Governor of Texas, “Governor Abbott Directs Comprehensive Data Center Audit,” press release, August 3, 2026; Office of the Governor of Texas, “Governor Abbott Directs TWDB To Penalize Data Center Reporting Failures,” press release, September 14, 2026.