You hear it all over the place now. Over coffee, over drinks, over and over, from Tories, from friends who are full-noise commie, from people with no interest in politics whatsoever.
What’s AI going to do to everything? What’s there going to be for our kids?
The political scientist Brian Klaas wrote a piece this week asking where the Apollo speech for AI is. Who’s going to stand up, he asks, and articulate a bold vision for what a good outcome for humanity might actually look like in a post-AI world?
Easy enough to say, a whole lot harder to do. How do we corral any of it? How do we get any say at all in what the tech bros have done to hurtle us towards a new and possibly very different tomorrow? How do you do an Apollo speech for that, when it feels as though this has been taken out of your hands, out of my hands, out of the control of any government?
I have some thoughts. But first, let me fill you with the fullest gloom of the thing. This is from a recent New Yorker piece:
If everything went right, the founders believed, artificial intelligence could usher in a post-scarcity utopia — automating grunt work, curing cancer, liberating people to enjoy lives of leisure and abundance. But if the technology went rogue, or fell into the wrong hands, the devastation could be total. Sam Altman wrote on his blog in 2015 that superhuman machine intelligence does not have to be the evil sci-fi version to kill us all. A more probable scenario, he suggested, is that it simply doesn’t care about us much either way — but in an effort to accomplish some other goal, wipes us out.
So. Utopia or extinction, and maybe no way to pay for your Netflix.
All most of us have ever asked for is to do a fair day’s work for a fair day’s pay. The sweep of history tells us that sometimes you will get that deal, and sometimes the table will be so rigged you’ll get crumbs, or chains.
Keynes predicted in 1930 that his grandchildren’s generation would work fifteen-hour weeks, prosperity having solved the economic problem. He was broadly right about the prosperity. He was completely wrong about where it went. Two earners working forty-plus hours a week, burnt out, time-poor, the gains captured by capital rather than converted into leisure. As Brian Klaas puts it, staggering prosperity tethered to surging alienation.
And boy have those prospects taken a more precarious turn in the last few years. Here’s your degree, and here’s the AI monster coming for your future.
What’s safe for today’s graduates? Law? Accounting? Medical diagnostics? Financial analysis? Coding? Anything?
Some of the conversations over coffee and drinks have been more along the lines of the machines getting brainier than us by far and then letting the robot dogs loose on us. Look, that’s absolutely part of the worry. But what I want to focus on is the more immediate question about how we all put food on the table.
For the most part I hear this expressed by parents of recent graduates describing the first, second and third jobs their very smart sons and daughters have bounced into, as the workplace incorporates the LLM elephant in every room.
The cardiologist I saw last year said: Oh, you’re a writer — what do you think about AI? He does all his own notes, won’t use it. He sees colleagues jumping in and he’s not comfortable about it. You hear it over and over.
In Miami this week, a woman who works in property development stepped up to the lectern of a graduating university class with the contented assurance such speakers tend to exude. She had not the slightest doubt that her good news would be the uplifting, inspiring message she considered it to be. She discovered almost immediately that they knew precisely of what she spoke, and they saw things very, very differently.
She happily declared that we are living in a time of profound change and that the rise of artificial intelligence is the next Industrial Revolution, the happy chirruping of an investor whose tulip stocks are doing well.
A sound of profound disapproval washed around the arena. A first wave, then a second, so loud that she realised she had hit a nerve of considerable sensitivity, though she expressed it with a nervous laugh. I struck a chord.
Oh no you didn’t.
May I finish? she said, and pressed on. Only a few years ago, AI was not a factor in our lives. That brought raucous ironic applause.
AI capabilities are in the palm of our hands, she said, before the crowd booed again. Oh I love it, passion: let’s go, she said gamely.
She tried to present the uncertainty of this moment as comparable to her own graduate encounter with the early days of the internet. But hoo boy, friend, a bar fight is not the same as the Battle of the Bulge.
She’s right that we’ve been through this before. It took us the better part of two lifetimes, from the first mills to the welfare state, to build the floor that caught people when the market dropped them. The people who could afford to wait out the Industrial Revolution’s adjustment period were not the people in the mills. The adjustment period is always paid for by the people with the least capacity to absorb it.
The students in that arena don’t have two lifetimes. The boos that rang through that arena echoed a very real anxiety for students everywhere. They are worried about how AI is already transforming some jobs and could eventually replace others altogether.
Easy enough to intone that this is all a great worry. But what the hell might we do about it? Is it just too big to tackle? Do we just have to hope that the billionaires over whom we have precious little control will do the right thing?
Or might there be, yet, some lessons from our recent past to give us some hope, or even guidance?
Let us consider Big Tobacco.
There’s no, er, smoking gun to point to here, but rather an accumulation over time of tactics and rules and laws that a coalition of health ministries, civil society organisations and developing country governments built into something they called the WHO Framework Convention on Tobacco Control. The point about this not-at-all-superhero-sounding name was that it gave them something they could point to in each country when the trade negotiators told them to forget about public health rules because there was a deal to be done. Thanks to that convention, they had the instrument they needed when it came time to reply, in bureaucratese, get fucked we will.
Consider also the legal action taken by Philip Morris against the sovereign nation of Uruguay. The cigarette company used an investment treaty clause to sue a small country for requiring plain packaging on cigarette boxes. Uruguay nearly buckled; it was only thanks to outside funding that they stayed in the fight. But in the end they won, and the real win for the movement wasn’t the verdict, but the wider effect. Health ministries across the world looked at what had happened and judged that they too could push harder.
Consider also that much of the momentum to fight back against Big Tobacco — to say: you will not sell it here; you will not sell it with lavish advertising: you will not sell it to our children — came from people rather than governments. Or at least not governments first. The motive power was ordinary people, with a specific grievance they could name, and enough organised fury to fight to win.
A sceptic could say: and yet hundreds of millions still smoke. But the extent to which the dial was moved, the extent to which a world industry was made to heel, is massive. They didn’t drive smoking out of the marketplace, but empire after empire that had once operated with free rein was compelled to constrain itself, country by country, on terms that are now very, very limited compared to what they once were.
That could be a model to think about here: internationally-developed and locally-enforced rules that oblige the AI companies to operate in certain ways or at least compel them to share the fruits of the enterprise.
Just as the Industrial Revolution was founded on slavery, this one is founded on stolen data. This invites fundamental questions about who might be owed what. I propose a radical approach to sharing, in the interest of the common good and the greater good.
I’ve written before about Universal Basic Services; the proposition that some things are too fundamental to human dignity and participation to be allocated purely by market mechanism. Healthcare and education are the obvious examples, though you could go further: transport, housing, food, broadband. You treat all of these as vital infrastructure, funded accordingly, the way we do roads and drainage but extended to other vital needs.
The AI moment makes this argument not more fanciful but more urgent. It also potentially makes it more fundable.
If the productivity gains are real, if AI genuinely stands to generate enormous wealth, then where could that wealth go, and where should it go?
How might you share it equitably and sustainably and for the betterment of humanity? More bluntly, how the hell do you find a way to tax and share a galaxy-sized golden goose?
Consider the companies who are building the AI systems, and using extraordinary amounts of computing power to do it. We’re talking about thousands of specialised chips running for months at a time, consuming electricity at a scale that would keep a small city going. That infrastructure has a footprint. It can be measured. It can be taxed.
How about a levy on the computing power used? A bit like a carbon tax, if you like, which doesn’t abolish driving, it just prices in the cost that driving imposes on everyone else.
The fund could be governed internationally; a sovereign wealth fund for the AI transition, seeded by the companies whose systems are doing the displacing and administered for the benefit of the people being displaced.
Enough? I’m not sure. It would need to cover an awful lot if the hype proves to be real. We’re not talking about automating one industry or one category of task. We’re talking about systems that stand to progressively do more of what lawyers do, what accountants do, what doctors do, what coders do, what analysts do, what writers do.
And then? All the wealth being generated in this new way makes its way into so very few hands. We have passed this way before and it took us two lifetimes to recognise that tax and a fair share are intrinsic to a fair and just society: From the few, back to the many. This would be the same principle that said the people who built the railways should not die in poverty while the railway owners built their second and third mansions.
Whether it might come as money, or as the goods and services that money would otherwise have bought — housing, healthcare, education, transport — is a detail of implementation. The principle would be the same: The scale of the sharing must match the scale of the taking.
I don’t see a levy like that as the whole answer so much as a mechanism pointing in a direction. The direction is: these gains were built on foundations that belong to everyone — public research, public infrastructure, and the data that all of us generated and none of us were paid for — and the proceeds should reflect that.
An adequate UBI is unaffordable, the critics like to say, and an affordable one is inadequate. Maybe. But Universal Basic Services sidesteps that argument because it isn’t a cash transfer. It’s a provision of the conditions for participation. A graduate who can’t find work in their trained field but has housing, healthcare, transport and retraining available isn’t in the same position as one who has none of those things and a debt clock running.
I’m not describing a totalitarian command economy. The ambitious people, the entrepreneurs, the investors would still be there, and the market wouldn’t disappear. It would just no longer be the only thing standing between you and destitution.
I know what this invites:
Bless, he thinks an international convention is going to do any good.
What’s he smoking?
That won’t get past go, won’t even collect two hundred dollars.
And I’m not saying this with anything like a lot of optimism or confidence. Who has ever managed to make a Thiel or Musk do anything they didn’t want to?
But I do know this: a movement large enough and riled enough, can surprise itself with its own strength.
We’re all in this together is history’s enduring lesson. It’s right to insist on a share.