Before We Worry About Being Bored in AI Utopia, Who Gets To Live There?

Apparently, one of the problems we may have to confront in the age of artificial intelligence is having too little to worry about. The machines will do the work. Disease will yield to scientific discovery. Our material needs will be met, leaving us with the terrible burden of figuring out what to do with ourselves.

I am willing to give it a shot.

More healthy years with the people I love, less anxiety about the future and some additional time on the water sound pretty good to me. I suspect I could keep myself occupied. But before we get too far into whether paradise might become tedious, I have a question: Who gets to live there?

That was my reaction to Nick Bostrom’s conversation with Spencer Klavan in The New York Times. Bostrom, whose book “Superintelligence” helped bring the dangers of advanced AI into public debate, is now also examining what happens if we get this right. His book “Deep Utopia” explores the meaning of human life in a world where technology has removed many of the constraints that shape it.

The headline presents this as a turn away from worrying about apocalypse toward worrying about utopia. Bostrom’s actual position is more careful. He says his outlook has not fundamentally changed. He still sees existential risks. He simply thinks the upside deserves serious examination, too.

Fair enough. We should have some idea of what we hope to accomplish with all this technology.

Then he offers the premise that stopped me: “Let’s imagine everybody gets a slice, everybody is free.”

I would like to spend a little more time on that sentence.

Who Owns the Solved World?

Bostrom knows what he is doing. His “solved world” is a philosophical thought experiment. He assumes technological maturity and sufficiently good governance so he can investigate what people would value after urgent material problems are addressed. He explicitly distinguishes controlling AI from deciding what humans use it for.

My concern is what happens when we carry that imagined destination into arguments about the technology being built today. The shared prosperity cannot remain an assumption while the machines become real.

An economy capable of producing extraordinary abundance still needs some arrangement for deciding who receives it. Ownership does not disappear because production becomes more efficient. Nor does the power that comes with ownership.

If AI and robots perform most economically productive work, the people who own those systems would have an obvious claim on the proceeds under familiar business arrangements. The position of everyone else is considerably less obvious. For people who earn their living by selling their time and skills, a reduced need for human labor raises a practical question about how they support themselves.

Being told that the economy can produce more than ever would be little comfort if you no longer had an income with which to buy any of it.

Lower prices could help enormously. So could tools that allow individuals and small businesses to accomplish things previously beyond their reach. I welcome both possibilities. But access to a capable AI assistant is not the same as owning a share of the productive economy. It does not automatically provide a home, healthcare or financial security.

Even if some form of distribution meets everyone’s basic needs, the terms matter. An enforceable right to a share of prosperity gives a person a different kind of security than an allowance someone else can withdraw. A comfortable life can still leave you dependent on decisions you have no meaningful ability to influence.

That is why freedom belongs in this discussion alongside abundance. Bostrom includes both in his premise. Anyone invoking a similar future to justify present decisions owes us an explanation of how both will be secured.

The Bills Arrive Before Utopia

There is also the rather inconvenient matter of getting from here to there.

A thought experiment can move directly to the point where the difficult arrangements have been settled. A family cannot. The mortgage, rent, groceries and healthcare expenses continue through whatever economic transition AI brings. Children do not postpone growing up until the labor market has adjusted.

We should be honest about what we do not know. It is possible that AI will create new kinds of work and expand opportunity in ways we cannot yet see. It is also possible that some people will lose livelihoods faster than alternatives become available. There is no reason to assume the benefits and disruptions will arrive in a convenient sequence, or that every community will experience them the same way.

Someone who loses a job at 55 cannot pay the bills with an assurance that future generations will be richer. Even if that assurance eventually proves correct, we still owe that person a serious answer about the intervening years.

The answer might involve several things: Income support, more affordable essential services, opportunities to own a stake in the businesses generating the wealth, and ways for people to participate in decisions affecting their livelihoods. None is simple. Each comes with costs, tradeoffs and questions about who exercises control.

Universal basic income is one proposed approach, but writing people a check would not settle every question. The amount, what it buys, how reliably it arrives and what happens to public services would all matter. So would whether recipients retain a real voice in the system supporting them.

I am not pretending to solve that here. I am asking that we give these questions attention proportional to the promises being made. If freedom from compulsory work is part of the attraction, the arrangements that make that freedom possible deserve more than a reassuring sentence about everybody getting a slice.

And that promise has to extend beyond the people fortunate enough to own the technology or live in the countries best positioned to benefit. “Everybody” is a very large commitment.

I Can Find Something To Do

Once we get past the question of material security, I am less troubled than either participant sometimes seems to be by the prospect of finding purpose.

At one point, Bostrom considers whether choosing furnishings would feel pointless if a system already knew your preferences and could do a better job. More broadly, why make an effort when a button could deliver the desired result?

Because sometimes making the effort is what you enjoy. Cooking dinner for someone can be satisfying even when a restaurant could produce a better meal. Playing an instrument does not require believing you are the best musician available. I do not need to be the most efficient person on the water to enjoy being on my boat.

Bostrom recognizes this, particularly when he describes a child making a birthday drawing for a parent. A machine might produce a more polished picture. It could not make that picture a gift the child chose to create. The relationship and the effort are part of what the parent values.

There is plenty of human purpose in that example. It survives without anyone having to invent a new species of consciousness.

I also understand that work can provide structure, friendship, pride and a sense of being useful. Losing those things can be devastating, even when money is available. A society in which fewer people need to work would have to take that loss seriously and develop other opportunities for contribution and belonging.

But I would not preserve economic insecurity to make sure people have a reason to get out of bed. Nor would I defend disease and physical decline as necessary ingredients of being human. When Bostrom describes having more healthy time to read, learn and explore, I find that appealing. We have enough good things to do without needing suffering to justify them.

The freedom to choose what those good things are would be essential. Some people would want to embrace every new capability. Others would want familiar lives with fewer burdens. People should be able to keep their privacy, decline an enhancement or do something themselves without losing their place in society.

Those choices cannot belong exclusively to the people who build the systems. Knowing how to make a machine more capable does not give anyone special authority to decide what the rest of us should want from life.

I am glad Bostrom is asking what success might look like. The AI conversation needs room for aspirations beyond preventing disaster and increasing output. Healthy time, meaningful relationships and greater freedom are good reasons to build better technology.

We should be able to ask for evidence that we are moving toward those goals. Are people becoming more secure? Do they have more control over their lives? Are the benefits reaching beyond the owners and early winners? Those questions should accompany every celebration of what the machines can now do.

Bostrom can imagine a world where everybody gets a slice and everybody is free. Making that world possible is a responsibility for the people building businesses, writing policy and exercising power now.

I’m willing to take my chances with being bored in a world where people are healthy, secure and free. First, tell me who gets to live there.