What Is Intelligence in the Age of AI?

Introduction

This fall, a teacher-education class gave its students a debate. Side 1: the people we call intelligent are strong in language and mathematics; beyond those, we are talking about coping and talent, not intelligence. Side 2: that view is too narrow; book smarts are not street smarts, and a truly intelligent person needs both.

Side 2 is closer, but neither side goes far enough. Both treat intelligence as a list of abilities and argue only about which abilities belong on the list. Neither asks what the abilities are for. My own answer is older than either side: intelligence grows out of wisdom, and wisdom begins with the fear of God.

Then I noticed that the loudest argument of our time, whether machines have reached “general intelligence,” is the same debate on a much larger stage. So this essay carries the classroom question into the age of AI. It looks at what the AGI debate reveals, pulls apart three words we tend to blur, asks what psychology says about wisdom and what Scripture adds, faces one hard fact, and ends with a question for parents: what is learning for?

1. The question under the AGI debate

Everyone seems to be arguing about AGI, artificial general intelligence. Is it already here? Is it five years away, or fifty? Will it take our jobs or solve our problems?

Under all of these sits an older question that rarely gets asked out loud: what is intelligence? We cannot say whether a machine has reached general intelligence if we cannot say what intelligence is. And we cannot.

This is not a new embarrassment. A standard textbook for teachers says plainly that psychologists have never agreed on a single definition.¹ Two AI researchers once gathered definitions of intelligence from psychologists, dictionaries, and their own field, and stopped at about seventy.² Melanie Mitchell, who studies both AI and human thinking, wrote in Science that what AGI means, and whether it means anything coherent, is still “hotly debated.”³

The AGI debate, in other words, is a symptom. It shows how unclear we are about the thing we are trying to build, and about the thing we are trying to grow in our kids.

2. Every test passed, and still no answer

The first intelligence test was not designed to find out what intelligence is. In 1904 Alfred Binet was asked to tell which kids could profit from regular classroom teaching and which needed special help, so the items that stayed in his test were the ones that predicted success in school.¹ Ever since, “intelligent” has quietly meant something close to “good at school”: strong with words, strong with numbers.

By that measure, the machines have arrived. In 2023 OpenAI reported that GPT-4 scored around the top 10 percent on a simulated bar exam and near the 90th percentile on the LSAT and SAT Math.⁴ A later study argued that the bar exam figure was inflated: against a more typical group of test takers, it falls to somewhere between the 48th and the 69th percentile.⁵ That is still a passing score.

Some researchers saw this coming. In 2019 François Chollet argued that a test score shows only a skill, and a skill can be bought with enough practice. Intelligence, he proposed, is how efficiently you acquire new skills in situations you have never seen.⁶ He built a set of puzzles, called ARC, that an ordinary person can solve without special training. When the second version came out in 2025, pure language models scored zero. By October 2026 the best system on the official leaderboard scored 95 percent. The average person scores around 60 to 66 percent.⁷

Even the test built to escape the old definition has been passed.

Yet the debate is not over. In a 2025 survey run by the Association for the Advancement of Artificial Intelligence, 76 percent of respondents said that scaling up current approaches is unlikely or very unlikely to produce AGI.⁸ That survey came before the ARC scores climbed, and views may have moved since. New tests keep being proposed. One group of 33 researchers suggested scoring AI systems on adapted human psychometric batteries, the descendants of Binet’s tests, and gave GPT-5 a score of 57 percent.⁹ None of these is accepted as settling the question.

There is a name for what keeps happening. The anthropologist Marilyn Strathern put Goodhart’s law this way: “When a measure becomes a target, it ceases to be a good measure.”¹⁰ Every time we write intelligence down and turn it into a test, the test becomes a target. Once it is hit, we find out the test was not intelligence after all.

That gap is the most useful thing the AGI debate has given us. A machine can pass the tests we wrote and still not be what we mean. So the tests were measuring something else all along: a sample of intelligence, chosen for a purpose. What we mean by intelligence was never fully written down.

3. Three words, and what AGI is defined to measure

It helps me to pull three words apart.

• Clever (or smart): getting results quickly in a particular environment.
• Intelligence: the ability to understand, reason, and solve problems. This is the part the tests sample.
• Wisdom: knowing what is good, and what our abilities are for.

The first two say what we can do. Neither says what for.

Picture a group whose rules reward betrayal over cooperation. The people who simply adapt to the rewards learn to betray. Robert Sternberg describes practical ability as using what you know to adapt to, shape, or select your environment.¹ The people who learn to betray have used only the first of the three. Shaping a world, or choosing a different one, means asking what the world ought to be like, and that is already a question of what for. Someone who only adapts to rewards may become clever, but not necessarily wise. AI research has a name for the same pattern, reward hacking: a system finds a way to score that its designers never intended.¹¹ The skill is real. The direction is missing.

Now look at AGI with these three words. OpenAI’s charter of 2018 defines AGI as “highly autonomous systems that outperform humans at most economically valuable work.”¹² Microsoft and OpenAI’s 2023 agreement reportedly went further and tied AGI to profit: OpenAI reaches AGI only when its systems can generate at least $100 billion in profits.¹³ When the two companies rewrote their partnership in October 2025, they named a panel, not a test: any AGI declaration by OpenAI would be verified by independent experts.¹⁴ Six months later a new agreement made the payments between them “independent of OpenAI’s technology progress” and did not mention AGI at all.¹⁵

Serious researchers disagree about when anything like AGI will come; Yoshua Bengio, for one, takes seriously the view that it could be only a few years away.¹⁶ I am not arguing about the date. I am asking what the definition measures. It does not measure understanding, and it does not measure wisdom. It measures economic return. In the terms above, AGI as industry defines it is a measure of clever: getting results that pay, in the environment we already have.

That is why it cannot answer our question. A definition built on what earns will find intelligence in whatever earns.

4. What psychology sees in wisdom, and what it does not say

Psychology has not ignored wisdom. Sternberg describes it as applying what we know, “as mediated by values,” toward a common good, balancing our own interests, other people’s, and the world’s.¹⁷ The teachers’ textbook puts it simply: using one’s abilities for the good of oneself and others.¹

The letter of James says something close.

Who is wise and understanding among you? Let him show by his good conduct that his deeds are done in gentleness of wisdom.James 3:13

But the wisdom that is from above is first pure, then peaceful, gentle, reasonable, full of mercy and good fruits, without partiality, and without hypocrisy.James 3:17

In both, wisdom shows in conduct, and it bears fruit for others.

Psychology has also studied how wisdom grows. Experience in the right settings matters more than age;¹⁸ adolescence is a key period;¹⁹ stepping back from our own concerns helps us reason more wisely.²⁰ What psychology does not say is where wisdom comes from in the deeper sense: which values are the right ones, and who decides what the common good is. Sternberg’s definition needs a good that it cannot supply by itself.

Here I should say where I stand. You do not have to share my faith to follow the rest of this essay, but you should know the ground it stands on; I have laid that ground out in The Foundations of Whole-Person Education. For me, Scripture speaks to exactly this question. “For Yahweh gives wisdom. Out of his mouth comes knowledge and understanding” (Proverbs 2:6). “But if any of you lacks wisdom, let him ask of God, who gives to all liberally and without reproach, and it will be given to him” (James 1:5). James even sorts wisdom by its source. One kind “comes down from above.” The other “is earthly, sensual, and demonic,” and “where jealousy and selfish ambition are, there is confusion and every evil deed” (James 3:15–16).

Psychology describes how wisdom grows. Scripture says where it comes from, and where it begins:

The fear of Yahweh is the beginning of wisdom. The knowledge of the Holy One is understanding.Proverbs 9:10

5. Why wisdom grows intelligence

Sternberg treats wisdom as a good use of intelligence: first the ability, then the wisdom to use it well. I believe the order runs the other way. Intelligence grows out of wisdom.

Two clarifications first. By wisdom I mean wisdom in the sense of Proverbs and James. It begins before any decision about purpose: it begins with the fear of Yahweh, and it includes awe, humility, and a delight in searching things out. And “grows out of” is about dependence, not a timeline. Knowledge and understanding also come from God (Proverbs 2:6), and the two grow together, like a spiral. But a mind bent by selfish ambition stops growing straight, and a mind that has stopped wondering stops reaching. Wisdom is the soil. I see five ways it feeds what grows in it.

Awe opens the door. Psychologists describe awe as meeting something so vast that our current understanding cannot hold it, so that we have to rebuild it.²¹ That rebuilding is learning. In one set of studies, adults who were led to feel awe became more aware of the gaps in their own knowledge, and that awareness predicted interest in science.²² In another, 444 kids aged 4 to 9 who looked at awe-inspiring scenes wanted to explore more and were more motivated to learn than kids who looked at ordinary scenes.²³ Research on awe in young kids is still thin, but it points the same way.

Watch a toddler meet something for the first time: a staircase, the ocean, the moon. There is a kind of reverence in the trying. The world is bigger than they are, and they know it. Scripture has words for that moment:

When I consider your heavens, the work of your fingers, the moon and the stars, which you have ordained; what is man, that you think of him?Psalm 8:3–4

Humility makes us teachable. Awe makes the self feel smaller, and that leads to humility.²⁴ Intellectual humility, knowing the limits of what we know, already shows up in kids aged 6 to 8, and there it goes together with intelligence.²⁵ Among the marks of the wisdom from above, James lists “reasonable” (3:17). The Greek word appears nowhere else in the New Testament; it describes someone willing to yield, open to being persuaded.²⁶ Persuaded by reasons, that is, not by pressure. A wise person can be taught. “The fear of Yahweh teaches wisdom. Before honor is humility” (Proverbs 15:33).

Wisdom clears away what bends the mind. James puts jealousy and selfish ambition at the root of the earthly kind of wisdom (3:14–16). Where there is self-interest, reasoning goes to work for it. Psychologists call this motivated reasoning: when we want a conclusion, we reason our way toward it.²⁷ Wisdom is not morality added on top of intelligence. It removes what bends intelligence out of shape.

Curiosity gives the engine. Curiosity helps memory, even for things we only meet in passing.²⁸ Curiosity and effort together predict school performance about as well as intelligence does.²⁹ In a study of 6,200 kids, those whose parents rated them as more curious scored higher in reading and math at kindergarten, and the link was stronger for kids from low-income families.³⁰ Young kids test hypotheses and draw causal conclusions much as scientists do.³¹ And when an adult shows them exactly how a toy works, they explore less and discover less.³²

The most fruitful curiosity often wants nothing back. A student who had learned his first theorem asked Euclid what he would get from it. Euclid turned to a servant: “Give him threepence, since he must make gain out of what he learns.”³³ Richard Feynman, burned out, decided to play with physics just for fun, and worked out how a tossed plate wobbles. Later he said the work that won him the Nobel Prize came from “that piddling around with the wobbling plate.”³⁴ G. H. Hardy wrote in 1940 that real mathematics is “harmless and innocent” and that no one had found any warlike use for the theory of numbers.³⁵ Less than forty years later, number theory became the basis of public-key encryption.³⁶ I am not saying useless questions pay off in the end, though they often do. The person asking was not asking for the payoff.

Scripture keeps awe and curiosity together, in a single psalm. “Yahweh’s works are great, pondered by all those who delight in them” (Psalm 111:2). The same psalm ends: “The fear of Yahweh is the beginning of wisdom” (111:10). James Clerk Maxwell had verse 2 carved, in Latin, on the doors of the Cavendish Laboratory in Cambridge.³⁷ And Proverbs says: “It is the glory of God to conceal a thing, but the glory of kings is to search out a matter” (25:2).

Wisdom gives the direction. Clever shrinks every problem to one question: how do I win? Wisdom widens it: who is this for? “Now the fruit of righteousness is sown in peace by those who make peace” (James 3:18).

None of these studies proves the order I am claiming. What they show is that the posture Scripture calls wisdom, awe and humility and open curiosity and care for others, is the posture in which learning happens. The order itself is a conviction, and I hold it as one. So wisdom gives intelligence an open door, a teachable heart, a clear mind, an engine, and a direction. That is what I mean when I say that intelligence grows out of wisdom.

6. A hard fact

There is a fact this essay has to face. A review of 63 studies found a modest negative correlation between religiosity and intelligence scores, about r = −.24 among college students and the general population, and the link was stronger for religious beliefs than for religious behavior.³⁸ On the tests’ own terms the correlation is there, and I am not disputing it.

What I want to point out is the layer it measures. Religiosity is a question about religion, not about faith. Religion is a human framework, with its own structures and its own problems, and most of those belong to another discussion. The surveys ask how strongly a person holds a religion’s beliefs, or how often they take part in its practices. James would not be surprised that belief by itself tells us little: “You believe that God is one. You do well. The demons also believe, and shudder” (James 2:19). For James, faith shows in works (2:18), and wisdom shows in conduct (3:13). The wisdom this essay is about is a posture before God, awe and humility and teachability, not an answer on a questionnaire. None of this explains the correlation away. It says that my claim sits at a different layer from the one the surveys measure.

The researchers offered three explanations.³⁸ Intelligent people may be less likely to conform, and so more likely to resist religious dogma. They may favor analytic over intuitive thinking. And religion may meet needs, such as a sense of control, self-regulation, and secure attachment, that intelligence can also meet. The third treats religion as a human framework serving human needs, which is what I mean by religion.

The first bears directly on this essay. Where religion works as power, deciding what may be asked, who may speak, and which answers count, it does bend the mind, just as self-interest does. That use of power is not biblical; Jesus himself condemned it. To the religious experts of his day he said: “Woe to you lawyers! For you took away the key of knowledge. You didn’t enter in yourselves, and those who were entering in, you hindered” (Luke 11:52). To his own followers: “But don’t you be called ‘Rabbi,’ for one is your teacher, the Christ, and all of you are brothers” (Matthew 23:8). In whole-person education, authority belongs to God, not to the teacher and not to the institution.

So I am not claiming that believers score higher on tests. I am claiming that intelligence has a source and a direction, and that wisdom is given, not earned: God “gives to all liberally” (James 1:5).

7. For parents: what is learning for?

If AI can do what the tests measure, and more and more it can, then a kid who learns only to score is training for a race against a machine that has already won. Many families still define smart the old way: grades, rankings, test scores. That definition is not wrong; it is small. On its own, it quietly trains kids to adapt to rewards, and it grows clever.

What is learning for? Our tagline says it in eight words: “Curiosity is the spark. Meaning gives learning direction.” This essay says the same thing at greater length. Learning is for growing in wisdom: awe at what is bigger than us, the humility to be taught, curiosity that wants nothing back, and abilities used for the good of others. Intelligence grows along with it.

A few places to start at home:

• Protect wonder. Take your kid somewhere bigger than they are, a night sky, a coastline, a mountain, a museum hall, and let them stand there without a worksheet.
• Let useless questions live. “Why does the plate wobble?” does not need a payoff.
• Don’t show them how too quickly. Ask what they think first. Let them try before you demonstrate.
• Watch what you praise. Praise the thing they made and the question they asked, not the rank.
• Make “I was wrong” a good sentence in your house, starting with yours.
• Ask who it is for. When your kid makes something, with AI or without it, ask not only “does it work?” but “who is it for?”

Our earlier essay, Learning in the age of AI is not just about skills, is about how a kid should work alongside a machine: inspect, doubt, modify, test. This one is about why. Both rest on the same old sentence:

The fear of Yahweh is the beginning of wisdom.Psalm 111:10

If I am right, it is the beginning of intelligence too.

Notes

English Scripture quotations are from the World English Bible (public domain). The Chinese edition quotes the Chinese Union Version. Sources are listed in their original language.