What are we for when machines do the work?

I started working at sixteen, as an apprentice at HP in Widen. For the next 26 years, through HP, Fujitsu and Cisco, my answer to "what do you do?" came with a company name attached. When I left Cisco at the end of 2016, that answer went with it, and I had to describe what I'm for without one.

I think a lot of people are about to face the same question, and most of them won't have chosen to.

AI and robots are getting cheaper and more capable fast enough that "what do humans do when machines do most of the doing?" has become a planning question. It belongs in leadership teams, in governments and at the kitchen table, and it belongs there this decade. Most of the public debate is still about income: who pays the bills when the jobs go. That matters. The harder question sits underneath it. When effort is no longer economically necessary, what are people for?

Jobs were never where most meaning lived

People define themselves through their jobs. That doesn't mean they get meaning from them. Gallup has surveyed employees around the world for years, and only around a fifth to a quarter say they are engaged at work. Since the industrial revolution, a lot of work has asked people to behave like machines: hit the number, close the ticket, stamp the widget. It's no surprise AI is good at that work. We made it machine-like.

So I don't accept that automation will rip meaning out of everyone's life. For many people, the job was a stand-in for meaning, or the way they paid for the parts of life that carried it.

That doesn't make losing a job harmless. The social psychologist Marie Jahoda showed that employment provides things a paycheck doesn't: a structure for the day and week, contact with people outside your family, a part in a shared goal, status and identity, and a reason to get up and go. She called these the latent functions of work. Replace the income and take away the job, and people still suffer, because those things are gone. Anne Case and Angus Deaton traced what happened in the US when work, community and status eroded together for people without college degrees. The damage showed up in deaths from suicide, drugs and alcohol.

Both things are true. Work is often unfulfilling, and it still holds a lot of lives together. If we automate the work, we have to deliberately rebuild what it was holding up.

Abundance is the floor

The optimistic case is that technology will collapse the cost of living. One target I find useful to think with: housing, food, healthcare and education for a family of four for about $1,000 a month. At that level, a modest income buys security and time. People get to choose what to do with their days instead of taking whatever pays.

I want that world. I also think it's a floor, a precondition for choice. John Maynard Keynes wrote in 1930 that once the economic problem is solved, humanity's permanent problem becomes how to use its freedom well. We're still working on that one. Basic-income pilots give a reality check: Finland's 2017–2018 trial found recipients reported better wellbeing and less stress. Security helps people feel better. It doesn't hand anyone a reason to get up in the morning.

There's also a practical gap. Even at $12,000 a year, people need a way to earn it, or someone needs to pay it, during the years when costs are still falling and skills are already being displaced. And the costs that matter most, housing and healthcare, are the ones most tied up in land, regulation and incumbents. They will be the slowest to fall.

The relevance problem

The hardest version of this question belongs to people who love their work.

Picture finally having the freedom to spend your days on physics or mathematics, the thing you always wanted to do. Then AI does it better, faster and cheaper. You're free, and you're no longer needed. If you aren't relevant in the work you love, the satisfaction can drain out of it.

One common answer is to treat that loss as a push: let the activity go and put your energy somewhere higher-value. That works for a hobby. I'm not convinced it works for a physicist, a translator, an illustrator or a doctor whose craft is part of who they are. Telling them to "move up" can sound like their life's work was a stepping stone. And it assumes everyone has a higher-value activity waiting. For many, the thing that disappears is the thing they felt most capable doing.

Edward Deci and Richard Ryan describe three basic psychological needs: autonomy, competence and relatedness. Most optimistic answers about AI speak to the first and the third: more choice, more time with people. Very few speak to competence, the feeling of being good at something that matters. AI hits that need directly.

Where competence goes next

My answer is judgment. When machines handle more of the execution, competence moves to the parts they can't own: choosing which problems matter, framing them well, weighing the options against what we value, and taking responsibility for the outcome. Judgment keeps people as the authors of outcomes. Without it, they become spectators of machine output.

This also changes how we should raise and teach people. Using AI to finish this week's homework is the same small goal with less learning. Using it to attack a problem that would have been out of reach is a different thing altogether. The useful question for parents and teachers is which problem is big enough to deserve this child's AI-amplified effort. Kids still need the experience of mastering something hard on their own, so I'd also keep some AI-free practice spaces, the way music and sport have always worked.

Meaning lives in process and people

People still play chess, even though software has beaten the best human players for years. People still paint and draw, even though image generators are faster. We do these things for what it's like to do them, and for the stories we tell each other afterwards. AI competes on outcomes. It doesn't compete on the experience of doing something, or on who you do it with.

I use two questions for any activity. Do I do this for the outcome or for the experience? And could AI do it better? If it's for the outcome and AI does it better, hand it over and review the result. If it's for the outcome and I do it better, that's where I lead, with judgment and responsibility. If it's for the experience, I protect it, whatever the benchmark says. And the things done for the experience that only people can do well, care, community and shared stories, are where I want to spend the time automation gives back.

That time won't defend itself. Attention-economy products are built to capture exactly the hours that automation frees up. More free time does not automatically become time with your kids or your parents. Someone has to decide that it will.

From my purpose to our purpose

We tend to talk about purpose as something individual: my calling, my passion, my path. I think that framing comes partly from industrial schooling and work, which trained us as separate units and measured us one by one. The meaning that lasts usually comes from shared projects: a family, a team, a band, a community that organises itself around something it cares about. One person can't do much. A group of people with a shared goal and good tools can do a great deal, and AI makes coordinating that group much cheaper.

That has a dark side. Movements can turn into tribes, and shared identity can be used against outsiders. Self-organising communities also tend to work best when their members are already well-resourced and chose to be there. Building the same thing in a fragmented neighbourhood is much harder work, and it's the work that matters most.

The transition is the real risk

So far, a lot of technology has separated people more than it has connected them. Without community, fear and uncertainty grow. And many people hear about an abundant future without seeing any path into it. From where they stand, it looks like billionaires becoming trillionaires while everyone else stays where they are.

Even the optimists usually concede that the good outcome depends on getting through the next 20 or 30 years. That's a whole working life. Telling someone that costs will fall eventually doesn't address what it feels like to live through the middle. The transition needs visible, local and inclusive pathways: bridging income, retraining that leads somewhere, places where people meet, and real mental-health support for young people. I'd also like us to measure more than employment and GDP. If we tracked purpose, loneliness, connection and agency alongside the economic numbers, we'd see much earlier whether this is working.

Purpose is infrastructure

I don't see purpose as a luxury that comes after abundance. It's part of what makes people resilient. Purpose is closely tied to happiness, it keeps people going when things are hard, and it makes them intrinsically motivated, so they learn what they need to learn. Psychologist William Damon's research on youth purpose found that only a minority of young people report a fully developed sense of purpose, and many are drifting. That should worry us more than any job statistic.

I find one distinction useful here. A passion is something you love doing. A purpose is something you love doing that helps other people. AI gives people with a purpose far more reach. People without one get more free time and less reason to use it.

I also don't think purpose is usually found by sitting and thinking about it. It gets built by committing to a concrete problem with a concrete group of people. Pick a problem, find the people, start small. Helping a young person do that is one of the most useful things a parent or mentor can do.

If you want a place to start, write down your answer to "what am I for?" without using your job title.