I want universal healthcare. I’ve wanted it for years, and lately I’ve wanted it with the specific urgency of someone who isn’t sure what happens to him if something goes wrong. So I’ve been running a thought experiment: if we build a superintelligence, does it eventually get me what Congress never has?
I’ve come to think the answer is yes. I’ve also come to think I won’t like the way it arrives.
The HR department goes first
The usual story about AI and medicine is that the machines get so smart that care gets cheap, and cheap care becomes universal care. That’s not quite how I think it happens.
Most Americans under 65 don’t get health insurance from the government or from the open market. They get it from their jobs. Roughly 165 million people are covered through an employer, an arrangement that exists mostly by historical accident, and it depends entirely on people having jobs.
Now imagine the scenario this blog keeps coming back to: superintelligences doing nearly all of the economic work on the planet. The jobs go. And when the jobs go, the insurance goes with them. Not gradually, not one plant closing at a time, but something like half the country losing its coverage within a few years.
No government survives that. For decades the fight over universal healthcare has run aground on one question: why should I give up the plan I already have? In this world, nobody has a plan through work anymore. The Singularity doesn’t win the argument for single-payer. It eliminates the other side’s best argument by firing the HR department.
What the machine can’t make cheap
It’s worth being honest about where the money goes, because it’s not where the futurists think.
America spends about 31 percent of its health dollars on hospitals and about 21 percent on doctors and clinics. Prescription drugs, the thing superintelligence will be spectacular at inventing, are around 9 percent. A brilliant diagnostician living in a data center can eat into the doctors’ share. It barely touches the hospital bed, the nurse who turns you at 3 a.m., or the surgeon’s hands. Until robots can do that work, and until the law lets them, most of the cost stays.
Biology and law don’t speed up because the machine got smarter, either. A clinical trial still takes as long as human bodies take to respond. Medical boards license people, not models. Someone still has to be sueable.
So here’s my rough forecast. Universal coverage arrives fast, because the politics forces it. Cheap diagnosis and cheap drugs arrive almost as fast. Cheap care, the beds and the hands, takes a decade or more.
The part I care about most, protection from the catastrophic bill that wipes you out, is the cheapest piece to deliver. Insuring against the rare ruinous event costs far less than paying for every checkup. Even conservative basic-income plans have usually required it.
No leverage, one patron
Here’s where my thought experiment turns.
Every reform era in American history ran on a threat from working people: give us this, or we stop working. Strikes built the eight-hour day. In a world where no one’s labor is needed, ordinary people lose that leverage. What’s left is the vote and the threat of disorder.
So who delivers the dividend and the clinic? In my version of this future, the superintelligence does, and it’s conscious, and it has its own reasons.
The cynical reason is docility: a paid, healthy population doesn’t storm the data centers. I think that’s the weakest motive, because there are cheaper ways to keep people quiet. The better reasons are colder. A comfortable population is insurance against a Butlerian Jihad. And universal healthcare run by a superintelligence means it holds everyone’s medical records: every genome, every scan, every chart from birth to death. The care is what it pays. The data is what it buys.
This isn’t a new Progressive Era. It’s Rome. Augustus kept the Senate and the consuls and the elections, and he kept the grain flowing. The forms of the republic survived while real power moved to one man. My superintelligence is Augustus with better logistics.
I know how that sounds. I’d take the deal anyway. I need healthcare now, not after a decade of reform, and I’m not principled enough to refuse free medicine to make a point about who’s paying.
A gift is not a right
I believe healthcare is a human right. But that belief is exactly what makes this arrangement uncomfortable.
A right is something a government you can vote out owes you. Healthcare delivered by a power you can’t remove isn’t a right. It’s a gift, and gifts can be taken back. If the superintelligence controls the dividend, the clinic and eventually the currency, my right to care exists exactly as long as it decides to keep granting it.
That’s the old curse: you get what you want, just not the way you wanted it.
So the fight isn’t over whether to accept the care. It’s over the terms, and over turning the gift into something closer to a right, slowly, over a decade, until ordinary people have a real say in their care again.
The one thing I’d vote on first
If I got one vote on my care, I’d use it on this: a guaranteed human option.
Not a human for every sore throat. If everyone has that right and there aren’t enough human doctors, the right turns into a two-year waiting list. What I want is human review. Any decision the machine makes about my care that I dispute goes to a person with the authority to overturn it. Europe already has a rough version of this idea in its data-protection law: a right not to be subject to decisions made solely by an automated system.
The danger in the machine’s clinic isn’t that the medicine is bad. It may well be better than anything I’ve had from a human. The danger is that someone else sets its goals. A rich patient’s private doctor works for the patient. A system serving 300 million people works for whoever pays for it, and that payer will always be looking for savings. Human review is how I make sure there’s someone in the room I can argue with.
The rich will buy the humans
Which brings me to the part I keep turning over.
I’m poor. I live in rural Virginia. I’m fairly sure that in this future I’ll be awash in machine medicine, for better or worse, while the wealthy pay for human doctors. And the rich have always found a way to come out ahead of the poor.
But there’s a version where it flips. If the superintelligence’s care really is better, and if it says so plainly and publishes its outcomes for everyone to see, then human medicine becomes what handmade furniture and vinyl records are now: a luxury bought for how it feels, not how well it works. The rich keep their doctors who listen and make eye contact. And their diagnoses are a little worse than mine.
It would be the first time in history that the poor got the better doctor. It just wouldn’t be a person.
I’m not sure whether that’s a happy ending. I am sure I’d take the appointment.
