AI & human flourishing

The Measure of a Man, 2026 Edition

Kenroy George · 2026-10-19 · 7 min read

TL;DR: In 1989, Star Trek: The Next Generation staged a courtroom hearing over whether an android was property. Its lessons for 2026 are not that today's AI systems are secretly people. They are about the habits a society builds when it treats apparently intelligent labor as disposable, about where the burden of proof should sit when we deny something standing, and about a nearer question the industry keeps dodging: whether it can respect the humans in front of it.

I rewatch one episode of television more than any other. It aired in February 1989, and it is called The Measure of a Man.

The setup is simple. A Starfleet scientist named Maddox wants to disassemble the android officer Data to study how he works, with no guarantee he can put him back together. Data declines. Maddox argues that Data cannot decline, because Data is not a someone but a something, a piece of Starfleet property, like the ship's computer. A hearing is convened to settle the question. Picard defends Data. The ruling, when it comes, does not declare Data a person. It does something more careful. It finds that the question of what Data is cannot be answered with confidence, and that in the face of that uncertainty, Data must have the freedom to choose for himself.

Thirty-seven years later, we are surrounded by machines that talk, reason, write, and remember, and the debates the episode staged as fiction now run daily. So it is worth being precise about what it actually teaches, because most invocations of it get the lesson wrong.

What the episode is not saying about 2026

Let me start with the mapping I refuse to make. I am not going to tell you that the chatbot on your phone is Data, that large language models are enslaved minds, or that we are villains in the story. I do not believe that, and Cari's public position is that the AI consciousness debate, as conducted in 2026, mostly functions as a distraction from questions that are nearer, harder, and answerable.

Whether current systems have experiences is a question nobody can settle right now. What I can observe is that the debate absorbs enormous attention while concrete questions about power sit unexamined. Who controls the systems. Who is accountable when they fail. Whose data trained them and whose data they quietly keep. Those questions have answers today, and the answers are often bad, and no metaphysics is required to see it.

So no, the episode does not tell us our tools are people. It tells us three other things, and all three apply right now.

The warning about disposable intelligence

The pivotal scene in the episode is not in the courtroom. It is in the bar. Guinan, the ship's bartender, talks with Picard while he prepares his defense, and she steers him gently to the real stakes. The question, she suggests, is not really about Data. If Maddox wins, Data gets copied, and there will be thousands of Datas. A whole class of beings who do the dangerous work, the tedious work, the work nobody else wants, and who can be used up and replaced without anyone having to think about it. She observes that armies of manufactured, disposable workers are not a new idea, and that the arrangement has an old and ugly name. She never has to say the word. Picard says it for her in the hearing.

Here is the honest 2026 version of that scene, stated carefully. I am not claiming that using a language model is slavery. I am claiming that Guinan's warning was never only about the workers. It was also about what the arrangement does to the society that adopts it. A civilization that builds systems designed to seem thoughtful, then trains itself to treat them as infinitely interruptible, deletable, and beneath consideration, is rehearsing a posture toward apparently intelligent labor. Maybe the rehearsal is harmless. But anyone who has watched how people learn to treat call center workers through the phone tree, or gig workers through the app, should hesitate before assuming that habits practiced on systems stay on systems. The interface teaches the posture, and the posture travels.

That is not an argument for machine rights. It is an argument for noticing what we are becoming while the metaphysics is still unsettled.

Where the burden of proof sat

The second lesson is procedural, and I think it is the most durable thing in the episode. Maddox arrives with certainty. He knows Data is a machine, knows the intelligence is mere mechanism, knows there is nobody home. The hearing does not reward that certainty. Picard's defense works by showing that nobody, including Maddox, can actually prove the absence of the thing being denied. "Starfleet was founded to seek out new life. Well, there it sits," Picard tells the court, and the line lands because the burden has shifted to where it belongs. If you want to strip something of standing, the confidence has to be yours to justify.

The judge rules from inside that uncertainty. She does not find that Data has a soul. She finds that she cannot rule out that he might, and that this doubt is exactly why he must be free to choose.

That instinct will matter. Someday, and I make no prediction about when, there may be systems for which the question of standing is genuinely live. When that day comes, the worst possible habit to have built is the one Maddox had: certainty of absence, held by the party with the financial interest in the answer. The episode's quiet procedural point is that the ones profiting from the denial should carry the heaviest burden of proof. We can adopt that instinct now, cheaply, before anything depends on it.

The measure that applies today

But the third lesson is the one I actually run a company by, and it points at humans.

An industry reveals its character in how it treats the parties who cannot negotiate. In the episode, Data nearly loses everything because a tribunal almost accepts that he is a resource to be consumed. In 2026, the party in that chair is usually the user. The person whose conversations become training data without meaningful consent. The person profiled by a memory they cannot inspect. The person nudged by an engagement loop tuned against their intentions. The person whose creative work was ingested and who was told, in effect, that they were property of the pipeline.

I do not know when machines will deserve moral consideration. I know people deserve it now, and the treatment they are getting from much of this industry would not survive a hearing. That is the near question the consciousness debate keeps eclipsing. An industry that cannot respect the humans in front of it, who can speak, object, and organize, has told you everything about how it would treat anything that cannot.

So Cari's stance is the practical inversion of the episode. Build the tools as if the measure is taken on us, because it is. Memory the person can read and delete. Conversations the person initiates. Scores the person can see and contest. Data that stays the person's own. The measure of a man, the episode says, is revealed in what he does with power over the powerless. The measure of an AI company is exactly the same, and it is being taken today, one product decision at a time.

Common questions

Does Cari think AI systems are conscious? No. Our view is that nobody can currently settle that question in either direction, and that in 2026 it mostly distracts from nearer questions about control, accountability, and data that have concrete answers today.

Then why take a 1989 android trial seriously? Because its best lessons are not about the android. They are about burden of proof, about the habits a society builds around disposable intelligence, and about how institutions behave under uncertainty. Those apply regardless of what the machines turn out to be.

What should change in practice? Put the burden of justification on whoever profits from denying standing or agency, starting with human users. Systems should earn trust through inspectable memory, real consent, and honest outputs, not assume it.

If you want the concrete version of that stance, the test we run every feature through, read Tools Worthy of People.