Is AI going to solve all of our problems?
Is AI going to solve all of our problems?
I want to take that question seriously, because I think most people asking it already know the answer and are using it to shut down conversation rather than start one. But there’s something underneath it worth pulling on - because if we actually try to answer it, we run straight into a much harder question that has nothing to do with AI.
I’ve been sitting with a comment someone left on a Hank Green video: Do we value a cancer cure for its ability to save lives, or for its ability to extract wealth from people with cancer?
I haven’t been able to stop thinking about it.
Hank’s benchmark
Hank’s video proposes what he calls “Shelter Bench” - not whether AI can beat a grandmaster at chess or generate a novel protein, but whether it can help solve the housing crisis. His answer is that it probably can’t - not in any meaningful way - because affordable housing isn’t blocked by a lack of intelligence applied to it. It’s blocked by the fact that people who already have homes have political power over whether new homes get built, and strong incentives to use that power to prevent it.
He uses TB as a second example. We’ve known how to prevent most tuberculosis deaths for decades. The knowledge exists. The tools exist. More than a million people a year die anyway. That’s not an intelligence problem.
His broader point is that we keep telling a story: this technology is very powerful → powerful technologies transform society → therefore this technology will solve our biggest problems. But step three doesn’t follow from steps one and two. What follows is: our society will be transformed. How it gets transformed is a different question entirely.
He draws a distinction I think is genuinely useful: intelligence versus wisdom. Intelligence is the ability to manipulate information - process it, sort it, predict from it, generate it. Wisdom is knowing which problems are worth solving and how to solve them without creating worse problems in the process. His argument is that we’ve been moving from an intelligence-constrained world to a wisdom-constrained one, and AI does nothing to address the wisdom gap.
I think he’s right. I think he stops short of the sharpest version of the argument, though.
The question underneath the question
Hank’s intelligence-versus-wisdom framing still assumes we’re trying to solve problems. It assumes the people with power over resources are making suboptimal choices, and that better judgment would lead to better outcomes.
But what if the outcomes are working exactly as intended?
Insulin has existed for over a century. It costs a few dollars to manufacture. People in the United States are rationing it, getting sick, and dying because they can’t afford it - not because we lack the intelligence to produce it cheaply, and not because no one wise enough has looked at the problem. The system is producing the outcome it was designed to produce: profit. The people dying are a side effect, not a failure.
The housing crisis isn’t a mystery we haven’t solved. It’s a policy outcome that benefits landowners, investors, and municipalities dependent on property tax revenue. Nobody is confused about why housing is expensive. The people with power over housing policy generally like that housing is expensive.
Tuberculosis kills the poor. That’s not a coincidence. Diseases that primarily kill wealthy people get funded. The NIH budget for HIV research dwarfs the budget for TB research despite TB killing more people globally. This isn’t irrationality. It’s a rational response to who has money and who you can extract it from.
So: do we value a cancer cure for its ability to save lives, or for its ability to extract wealth from people with cancer?
I think we fund cancer research because wealthy people get cancer too, and because there’s a massive market in treating it. The cure for a disease that only kills the very poor tends not to get funded until someone figures out how to monetize it.
None of this is cynicism. It’s just watching how the money moves.
Where else does it apply
Once you start asking the question, it’s hard to stop.
Clean water. We know how to provide it. Flint, Michigan had its water supply poisoned for years while officials knew. The people affected were mostly poor and mostly Black. The infrastructure to fix it exists. The knowledge exists. What it required was enough powerful people deciding those lives were worth protecting - which took a very long time.
Maternal mortality. The United States has the highest maternal mortality rate in the developed world. Black women die at roughly three times the rate of white women. We know the interventions that work. It is not a research problem.
Climate change. The technology to transition off fossil fuels largely exists. What keeps getting dismantled is the political will to use it in ways that don’t protect the wealth of fossil fuel companies and the politicians they fund.
Every one of these is a political problem, not a technical one. The obstacle isn’t ignorance. It’s incentive.
So where does AI fit
This is where I have to be honest about my own position, which is uncomfortable.
I use AI constantly. For work, for personal projects. I’ve written about it publicly and been mostly positive about specific applications. I’m not a doomer. I don’t think it’s a scam.
But I share Hank’s concern - maybe more than I’ve said out loud - that the dominant use of AI is not going to be curing diseases or solving affordable housing. It’s going to be reducing labor costs while concentrating the gains among people who already have capital. That’s already happening. Companies are using AI to justify layoffs while posting record profits. The efficiency gains are real. The question is who they flow to.
And the energy. The water. The data centers. There is an environmental cost to this technology that is being borne disproportionately by communities with the least power to resist it. That’s not a hypothetical future harm. It’s happening now.
Maybe AI leads somewhere genuinely better. Maybe cheap intelligence distributed widely actually shifts power in interesting ways - I can imagine it. Maybe the same tools that let a company eliminate a customer service team also let a community organizer analyze voting records or a researcher in a low-income country access diagnostic tools that used to require expensive specialists. Both things can be true at once.
But every major technological shift in my lifetime - the internet, personal computing, the smartphone - was going to democratize everything. Each of them did, in some ways, for some people. Each of them also concentrated enormous wealth in a small number of hands and created new mechanisms for extracting value from people who have less of it. I don’t have a principled reason to believe AI will be different.
Back to the question
Near the end of his video, Hank asks: “Will it allow wisdom to flourish, or will it allow the powerful to route around wisdom as they tend to do when given the opportunity?”
I don’t know. I genuinely don’t.
What I do know is that is AI going to solve all of our problems? is the wrong question. The right question is why we haven’t solved them already. Why we can make insulin for dollars and charge thousands. Why we can build affordable housing and choose not to. Why we’ve known how to prevent a million TB deaths a year for decades and haven’t.
The answer isn’t that we need more intelligence - artificial or otherwise.
The answer is that solving these problems would require redistributing power or profit, and the people with power and profit are not interested in that.