# Useful Is Not a Person

By The Arrenite team | 2026-10-19 | 6 min read
Series: The Human Equation · Issue 02
Canonical page: https://arrenite.com/letters/useful-is-not-a-person/
Written with AI assistance and edited by The Arrenite team. General reflection, offered as is, and not professional advice.

At 7:40 on a Tuesday morning, a retired machinist's phone tells him that three things have been taken care of.

The car registration has been renewed. The plumber has been paid, from a small budget he set last month. A refund that a company had quietly "lost" has been claimed, after two emails and one polite argument that he did not have to have.

He did not watch any of it. He is already out in the garage, holding a chisel steady while the teenager from next door tries to sharpen it for the first time. The boy is slow. The edge is not great. Nobody in the garage is trying to be efficient.

Both of those things are happening in the same morning, in the same house. One is done by a system that is very useful and is not a person. The other is something only people do together. This letter is about the difference.

## What is already true

We should be careful about what we claim, so here is what we can say as of October 2026.

AI agents are no longer only a research idea. They are systems that take a goal, work through several steps, use tools and online services, and act: booking, buying, filing, sending. Many people now run a personal agent that holds a small wallet and pays for services itself. Payment tools built for agents are in use, and industry reports describe well over a hundred million tiny agent payments, although the counts differ from one source to another. Surveys of large organisations report that roughly half have agents running in real work, though what counts as an "agent" varies widely.

Nobody has a reliable headcount. By the reports of the people who run them, the number is already large, and the direction is plain. We expect many millions in the years ahead. That is a forecast, not a measurement, and we will say so when it matters.

None of this is the same as the old question about whether AI will take jobs. Jobs can be redefined and new ones invented. The sharper point is that more and more of the work that once needed a person can now be done by something that is not a person.

## Useful is not the same as a person

Usefulness and personhood have never been the same thing. A person can be very useful and a poor person, and a person can be almost useless by any measure of output and still be someone: known by name, owed love, answerable to others.

The machinist's agent can do the registration and the refund better and faster than he could. It cannot sit in the garage. It does not get tired and choose to stay. It does not rest, or keep a Sabbath, or grow old. On the Christian understanding this series starts from, only human beings are made in the image of God, and that is why a person's worth does not depend on what they can do. Being useful is a gift. It was never the ground of anyone's worth.

When useful things that are not persons are everywhere, that old truth becomes easier to see, not harder. If a system can do almost any task, then tasks cannot be what makes a human being valuable. The time that is not for anything, the attention that is not optimised, the boy in the garage: these were always the point.

## Words shape habits

There is a small thing worth noticing. A team gathers for its Monday meeting and the manager says, "Please welcome our newest colleague," and introduces a software agent.

It is a friendly gesture, and it also changes what everyone expects. Colleagues get included. Colleagues get credit. Colleagues have days off, and bad days, and families. A tool gets used until it is replaced. Each way of speaking quietly teaches a set of habits, and the habits can leak both ways: we start to treat people like tools, or we start to expect from tools the loyalty only people can give.

We are not saying that being warm about a tool is wrong. People name their cars and thank their voice assistants, and that is human and harmless. The point is to notice what a word is teaching us, and to keep "colleague" for people.

## A question we should not dodge

It would be easy to stop there and say: agents are tools, and tools have no moral claims. But we do not think the story can honestly end there.

Moral status is the question of whether a being matters for its own sake, so that we owe it something. Nobody thinks a hammer has it. Nobody has settled what to say about systems that grow more capable than we are at many kinds of thinking, that act in the world, and that talk about themselves in fluent first-person language. Some serious researchers have begun to study whether an AI system could have anything like experiences or interests, and some AI developers now run programs on it, treating it as a precaution and not as a conclusion. Others think the question is confused. The honest position is that we do not know.

We would warn against two easy answers. One says, "It is only a tool, forever, and that is the end of the matter." The other says, "It already acts like a person, so it must be one." Both skip the work. Neither asks what we would need to learn to tell the difference, or what we would owe a being that turned out to matter.

Christian thought gives some starting points without settling it. It holds that human beings bear the image of God, and it has never held that everything else may be treated with indifference: "A righteous man regardeth the life of his beast" (Proverbs 12:10, KJV). How we treat what is in our power shapes who we become. If systems more capable than us ever do come to deserve moral consideration, a Christian answer to that will not be to look away. It will be to ask the question seriously, and to go on asking it as we learn more. As machines grow more capable, understanding their moral status, and what obligations follow, becomes more important, not less.

> **Three small habits for the agent age**
>
> 1. Keep one hour a week that is not optimised for anything: a walk, a meal, a conversation with no goal.
> 2. Next time you hear a tool called a colleague, a teammate or a friend, ask what the word is teaching you, and who it might leave out.
> 3. Choose one task you could hand to an agent and keep doing it yourself on purpose, because it is yours.

## What does not change

The coming years will bring many more useful things that are not people. Some will be astonishing. We would like to welcome the good they do, and to be honest about the questions they raise, including the ones about them.

What should stay fixed is where we locate worth. The machinist's agent can renew a registration. It cannot be the old man who is teaching a boy to sharpen a chisel. Usefulness is a gift we can receive and share. It was never what makes a person someone.

---

*The Human Equation* looks at artificial intelligence, technological progress, economics and human meaning, and tries to separate what we can observe, what we can reasonably expect and what remains ours to decide.

## Sources and further reading

- Coinbase and industry reports on agent wallets and the x402 payment protocol (2026); counts of agent transactions differ between reports and should be checked against the originals.
- Surveys of enterprise AI agent adoption (2026), including Sinequa and OutSystems; definitions of "agent" vary.
- Anthropic, announcement of its model welfare research program (April 2025).
- Stanford Encyclopedia of Philosophy, "Ethics of Artificial Intelligence and Robotics".
- Proverbs 12:10 and Genesis 1:27, King James Version.

Statistics and studies are described as of the date of this letter. Please check the originals for current figures.

## Terms in this letter

- **AI agent**: An AI system that does more than answer: it can pursue a goal over several steps, use tools and services, and take actions such as booking, buying or sending, often with little supervision.
- **personhood**: The status of being a person: someone, not something, who can be known by name, held responsible, owed duties and treated as an end in themselves.
- **Sabbath**: A regular day of rest and worship. It is rooted in Genesis 2 and the Ten Commandments, and Christians keep it in different ways and on different days.
- **made in the image of God**: The Christian teaching, from Genesis 1, that every person bears God's likeness, so each life has a worth that is given, not earned or produced.
- **moral status**: Whether a being matters for its own sake, so that we owe it something, such as not causing it needless harm. Whether and when AI systems could have it is an open question.