AI Can Take the Task. It Shouldn't Take Who You Are.

In The Economist this week, the Bartleby column takes up a question most AI coverage skips. In If AI displaces human workers, they will lose more than just pay, the column argues that even if society solved the problem of replacing lost income, it would still face a harder one. As its subtitle puts it, "Work has value beyond the income it provides."

Bartleby leans on a recent research review from Google DeepMind, Work, Wellbeing, and Choice: Empirical Lessons for AI Futures, and on the psychologist Marie Jahoda, who identified five hidden benefits of a job: structure for the day, contact with other people, a shared purpose, status, and a push toward activity. The column notes that most Swedish lottery winners kept working, and that long-term unemployed Germans reported a big jump in wellbeing once they reached retirement age, as if they had finally moved into a category society respects. Its conclusion is sober. Funding people through an AI disruption would be hard. Giving them a sense of self-worth would be just as hard.

I agree. I also think the column stops one step short of the real issue.

The Job Was Never Supposed to Be the Whole Identity

Bartleby frames this as a problem of dignity. I see it as a problem of identity, and specifically of where people have been taught to locate theirs.

For generations, the first question at any gathering has been "What do you do?" We answer with a title and an employer. Over time, many of us stop separating the two. The job becomes the story we tell about ourselves, and the organization quietly becomes the author of it.

That arrangement works until the job changes. Then people discover that the structure, the status and the sense of purpose they relied on were on loan. The German retirees in the research make the point vividly. Their circumstances barely changed at retirement. What changed was the label, and the label was carrying more of their self-worth than anyone realized.

I made a related argument after 60 Minutes looked at AI learning the jobs of experts, in AI Can Learn Your Expertise. It Can't Learn Who You Are. A model can absorb what you know. It can't absorb why you notice what you notice, or the problems you keep returning to whether or not anyone pays you to solve them.

Choice Is the Variable That Matters

The DeepMind review offers a detail worth more attention than it usually gets. Among the factors that shape how leaving work affects wellbeing, the authors put agency and choice first: whether stepping away was voluntary, and how much control people have over what comes next.

That finding speaks directly to conviction, which I describe as the earned capacity to act on what you believe before the outcome is certain. People who know what they naturally solve for have something to act on when a role disappears. They can choose a next chapter rather than wait to be assigned one. People whose identity lives entirely inside a job description have nothing to choose from, because the description was the whole of it.

None of this means jobs don't matter. The column is right that nothing else yet delivers the same bundle of income, routine and respect in such a socially acceptable form. Organizations provide real structure, and good ones should keep doing it well. Predictable rhythms, clear roles and shared goals help people do their best work. The goal is balance, where the institution provides structure and the individual brings an identity that doesn't depend on it.

What Leaders and Professionals Can Do Now

Bartleby also notes research finding that even an hour a week of volunteering improved retirees' sense of purpose. Meaning, in other words, doesn't require a full-time role. It requires a place where people can use what is distinctly theirs and be seen doing it. The social side matters too. I wrote about how much we lose when we stop gathering in We Stopped Hosting Each Other. We Lost More Than Dinner Parties.

A few questions to work through this week:

For leaders redesigning roles around AI: when you tell people their tasks are changing, do you also help them name what they bring that the tasks never captured? If not, you are asking them to give up part of their identity without offering anything to replace it.

For professionals: if your title disappeared tomorrow, how would you describe your value in one sentence without mentioning your employer? If that sentence is hard to write, start writing it now, while you still have the choice. And remember that the answer will keep evolving as you do.

For anyone who manages people: where do your people get structure, contact and purpose besides their task list? If the answer is nowhere, AI isn't the only risk to their wellbeing.

The column ends by saying self-worth would be as hard to provide as income. I'd put it slightly differently. Self-worth was never something an employer could hand out. It is built by people who know who they are, in organizations wise enough to make room for it.

AI may take over much of what we do, but only we can decide who we are while we do it.

Want to explore these ideas further? Learn more about my work on leadership, identity and conviction at www.theglennllopis.com.