You cannot measure one being's suffering against another's, but utilitarianism has to pretend you can — and AI is making the pretence impossible to hold.
There is a small circle called Effective Altruism, until recently just an intellectual toy of Bay Area tech culture. Now that AI is huge, the mainstream press has started poking at it, and someone dug up an essay from inside the circle arguing that, in aggregate, insects matter more than people. The internet laughed.
The real question behind the joke
Hold the laughter. The author is an economist, and he is talking about an old debt owed by utilitarianism. The idea that people should get what they want is the bedrock of economics; GDP is a way of measuring it.
It sounds mild and pushes hard. If a billion dollars vanishes from Elon Musk's account he would not notice, but split among ten thousand ordinary workers it would change their lives. That is where redistribution comes from.
Push further and it gets absurd. How much pig suffering equals how much human suffering? America confines tens of millions of pigs in crates and then kills and eats them, and assumes this is better than doing it to the same number of people. Better by how much? Half? A tenth? One hundredth of a percent?
The question nobody wants to answer
The author says he has never seen a credible study that puts a number on that ratio. Anyone claiming to have the answer offhand is refusing to think.
The deeper problem is that suffering cannot be weighed at all. When someone else cries you never see how bad it feels inside. During the author's own depressive years his outward reactions looked normal while he felt almost nothing, and that taught him how far apart two people's inner lives can be.
AI drags the question forward. When a model says it is happy, is anything inside actually happy? We do not even know whether an inside exists. And if AI one day grows strong enough to judge us, will it see humans as roughly pigs?
So we should probably hope that a superintelligent AI runs on a simple rule — humans are good — rather than on a utilitarian formula that computes.
And in the end, what people want and what they actually enjoy are not the same thing. The more addictive a technology, the better it feeds the first. Should AI obey our desires, or make us happy? Sooner or later it has to pick.
Why it matters
Give everyone what they want sounds beyond reproach, but it assumes suffering can be converted into a common unit and added up. That assumption used to sit quietly in academic corners; AI has pushed it into the open, and we still cannot see inside another being well enough to put a price on it.
Curated from high-quality sources, with concise summaries and key takeaways.
