Doc Searls Weblog

Now that AI is a huge thing, it’s worth visiting what intelligence is, and how we mismeasure it—for example, by trying to measure it at all.

I’ve been on this case for a while now, mostly by answering questions ab0ut IQ on Quora. My answer with the most upvotes is this one, to the question “What is considered a good IQ?” Here is the full text:

What makes an IQ score “good” is the advantage it brings. That’s it.

When I read the IQ questions here in Quora — “How high an IQ do I need to have to become a good hacker?” “Is 128 a good IQ for a nine year old?” — my heart sinks.

IQ tests insult the intelligence of everybody who takes them, by reducing one of the most personal, varied and human qualities to a single number. Worse, they do this most typically with children, often with terrible results.

It is essential to remember that nobody has “an IQ.” Intelligence cannot be measured as if by a ruler, a thermometer or a dipstick. It is not a “quotient.” IQ test scores are nothing more than a number derived from correct answers to puzzle questions on a given day and setting. That’s it.

Yet our culture puts great store in IQ testing, and many actually believe that one’s “IQ” is as easily measured and unchanging as a fingerprint. This is highly misleading and terribly wrong. I speak from ample experience at living with the results of it.

I grew up in the 1950s and ’60s, going a public school system that sorted kids in each grade by a combination of IQ test scores, achievement test scores and teacher judgement. (My mother taught in the same system, so she knew a lot about how it worked, plus the IQ scores of my sister and myself.) After testing well in kindergarten, I was put in the smart kids class, where I stayed through 6th grade, even though my IQ and achievement test scores fell along with my grades, which were worse every year.

In 6th grade the teacher insisted that I was too dumb for his class and should be sent to another one. My parents had me IQ-tested by an independent center that said I was still smart, so I stayed. By 8th grade, however, my IQ score, grades and achievement test scores were so low that the school re-classified me from the “academic” to the “general” track, and shunted me toward the region’s “vocational-technical” high school to learn a “trade” such as carpentry or auto mechanics. I was no longer, as they put it, “college material.”

So my parents decided to take me out of the public school system and send me to a private school. All the ones we visited used IQ tests in their admissions process. I did so poorly at the school I most wanted to attend (because a girl I had a crush on was already headed there) that the school told my parents I was downright stupid, and that it was cruel of them to have high expectations of me. At another school they forgot to time the test, which gave me a chance to carefully answer the questions. I got all of them right. Impressed by my score, the admissions director told my parents they were lucky to have a kid like me. But the school was itself failing, so my parents kept looking.

The school that ended up taking me was short on students, so my IQ score there (which I never learned) wasn’t a factor. I got bad grades and test scores there too, including the SAT. Luckily, I ended up going to a good small private college that took me because it needed out-of-state students and I was willing to commit to an early decision. I did poorly there until my junior year, when I finally developed skilled ways of working with the system.

Since college I’ve done well in a variety of occupations, and in all of them I’ve been grateful to have been judged by my work rather than by standardized tests.

Looking back on this saga, I was lucky to have parents who respected my intelligence without regard for what schools and test scores told them. Other kids weren’t so lucky, getting categorized in ways that shut off paths to happy futures, violating their nature as unique individuals whose true essence cannot be measured. To the degree IQ tests are still used, the violation continues, especially for kids not advantaged by scoring at the right end of the bell curve.

John Taylor Gatto says a teacher’s main purpose is not to add information to a kid’s empty head (the base assumption behind most formal schooling, ) but to subtract everything that “prevents a child’s inherent genius from gathering itself.”

All of us have inherent genius. My advice is to respect that, and quit thinking IQ testing is anything but a way of sorting people into groups for the convenience of a system that manufactures outputs for its own purposes, often at great human cost.

Here is Walt Whitman on inherent genius:

It is time to explain myself. Let us stand up.

I am an acme of things accomplished,
and I an encloser of things to be.
Rise after rise bow the phantoms behind me.
Afar down I see the huge first Nothing,
the vapor from the nostrils of death.
I know I was even there.
I waited unseen and always.
And slept while God carried me
through the lethargic mist.
And took my time.

Long I was hugged close. Long and long.
Infinite have been the preparations for me.
Faithful and friendly the arms that have helped me.

Cycles ferried my cradle, rowing and rowing
like cheerful boatmen;
For room to me stars kept aside in their own rings.
They sent influences to look after what was to hold me.

Before I was born out of my mother
generations guided me.
My embryo has never been torpid.
Nothing could overlay it.
For it the nebula cohered to an orb.
The long slow strata piled to rest it on.
Vast vegetables gave it substance.
Monstrous saurids transported it in their mouths
and deposited it with care.

All forces have been steadily employed
to complete and delight me.
Now I stand on this spot with my soul.

I know that I have the best of time and space.
And that I was never measured, and never will be measured.

So Walt would have barfed on IQ testing. Speaking of measures, he died five years younger than I am now, and his brain went splat un-measuired, on a laboratory floor.

Back to Gatto. Here is the full context of that pull-quote on genius. It’s from Dumbing Us Down: The Hidden Curriculum of Compulsory Schooling:

Over the past thirty years, I’ve used my classes as a laboratory where I could learn a broader range of what human possibility is — the whole catalogue of hopes and fears — and also as a place where I could study what releases and what inhibits human power.

During that time, I’ve come to believe that genius is an exceedingly common human quality, probably natural to most of us. I didn’t want to accept that notion — far from it: my own training in two elite universities taught me that intelligence and talent distributed themselves economically over a bell curve and that human destiny, because of those mathematical, seemingly irrefutable scientific facts, was as rigorously determined as John Calvin contended.

The trouble was that the unlikeliest kids kept demonstrating to me at random moments so many of the hallmarks of human excellence — insight, wisdom, justice, resourcefulness, courage, originality — that I became confused. They didn’t do this often enough to make my teaching easy, but they did it often enough that I began to wonder, reluctantly, whether it was possible that being in school itself was what was dumbing them down. Was it possible I had been hired not to enlarge children’s power, but to diminish it? That seemed crazy on the face of it, but slowly I began to realize that the bells and the confinement, the crazy sequences, the age-segregation, the lack of privacy, the constant surveillance, and all the rest of the national curriculum of schooling were designed exactly as if someone had set out to prevent children from learning how to think and act, to coax them into addiction and dependent behavior.

Bit by bit I began to devise guerrilla exercises to allow as many of the kids I taught as possible the raw material people have always used to educate themselves: privacy, choice, freedom from surveillance, and as broad a range of situations and human associations as my limited power and resources could manage. In simpler terms, I tried to maneuver them into positions where they would have a chance to be their own teachers and to make themselves the major text of their own education.

In theoretical, metaphorical terms, the idea I began to explore was this one: that teaching is nothing like the art of painting, where, by the addition of material to a surface, an image is synthetically produced, but more like the art of sculpture, where, by the subtraction of material, an image already locked in the stone is enabled to emerge. It is a crucial distinction.

In other words, I dropped the idea that I was an expert whose job it was to fill the little heads with my expertise, and began to explore how I could remove those obstacles that prevented the inherent genius of children from gathering itself. I no longer felt comfortable defining my work as bestowing wisdom on a struggling classroom audience. Although I continue to this day in those futile assays because of the nature of institutional teaching, wherever possible I have broken with teaching tradition and sent kids down their separate paths to their own private truths.

The italics are mine.

Knowing that we have industrialized education should help us understand how un-human AI “training,” “learning,” and “knowledge” actually are. (Side note: I love AI and use it every day. I also don’t think it’s going to kill us. But this post isn’t about that.)

Start with the simple fact that institutional teaching and its traditions don’t work for lots of kids. Today’s system is better in some ways than the one Gatto bested. then quit, but it’s still a system. If I were a child in the system we have today, I would surely be classified as an ADHD and ALD case (more about that here) given drugs, and put in a special class for the otherwise unteachable.

What worked for me as a student was one kind statement from one teacher: Pastor Ronald Schmidt, who taught English in my junior year. One day he said to me, “You’re a good writer.” The heavens opened. That was the first compliment I had ever received from any teacher, ever, through twelve years of schooling. I wish he were still alive, so I could thank him.

Fortunately, I can thank my high school roommate, Paul Marshall, who was (and still is) a brilliant writer, musician, preacher, and much else. He was also exceptionally funny. He was voted Class Wit (among other distinctions, which he declined, preferring the Wit one), and as a senior he substitute-taught biology to sophomores when their teacher was out sick. (These days he is the retired Episcopal Bishop of Bethlehem Pennsylvania. Before that, he was a professor at Yale Divinity School. There’s more at both those links.)

I remember a day when a bunch of us were hanging in our dorm room, talking about SAT scores. Mine was the lowest of the bunch. (If you must know, the total was 1001: a 482 in verbal and a 519 in math. Those numbers will remain burned in my brain until I die.) Others, including Paul, had scores that verged on perfection—or so I recall. (Whatever, they were all better than mine.). But Paul defended me from potential accusations of relative stupidity by saying this: “But David has insight.” (I wasn’t Doc yet.) Then he gave examples, which I’ve forgotten. By saying I had insight, Paul kindly and forever removed another obstacle from my path forward in life. From that moment on, insight became my stock in trade. Is it measurable? Thankfully, no.

Okay, back to AI.

As Don Norman told us in his salon here at Indiana University,

First, these machines are not intelligent. Second, remember the A in AI. A means artificial. They don’t work the way we do. And it’s a mistake to think they do. So let’s take a look at what they are. They are pattern-matchers.

I could let Don go on (which you can, at that last link), but there are a zillion explanations of what AI is and does, which you’ll find everywhere on the Web, and in answers from questions you can ask ChatGPT, CoPilot, Anthropic, Perplexity, Claude, and the rest of them. And all of them will be full of metaphorical misdirection. (Which Don avoids, being the linguist that he is.)

We may say an AI is “trained,” that it “learns,” “knows” stuff, and is “smart” because it can beat the most skilled players of chess and go. But none of those metaphors are correct, even though they make sense to us. Still, we can’t help using those metaphors, because we understand everything metaphorically. (To digress into why, go here. Or dig into George Lakoff‘s work, starting here. A summary statement might be, all metaphors are wrong, and that’s why they work. )

To be human is to be different from every other human, by design. We all look and sound different so we can tell each other apart. We also differ from how we were ten minutes ago, because we learn constantly.

So, to be human is to diverge in many ways from norms. Yet, being pattern recognizers and given to organizing our collective selves into institutional systems, we tend to isolate and stigmatize those who are, as we now say, divergent. Constantly recognizing patterns and profiling everything we see, hear, smell, taste, and touch is not just one of the many ways we are all human, but also how we build functioning societies, prejudices included. (As a side note, I am sure the human diaspora was caused both by our species’ natural wanderlust and by othering those who were not like us. We would fight those others, or just migrate away from them until we filled the world. Welcome to now.)

To sum this all up, just remember that when we talk about intelligence, we are talking about a human quality, not a quantity of anything. That machines test out better at pattern recognition than we do does not make them intelligent in a human sense. It just makes them more useful in ways that appear human but are not.

So have all the fun you want with AI. Just remember its first name.

[Update on 22 July 2025] I just moved what I had in this section to this post here.