I believe that AI can have the same scale of positive impact on human flourishing as low-cost energy—but only if we properly value it, integrate it, and liberate it.
This is what I argue in my first ever speech on AI—which I delivered earlier this month at my annual private energy conference. The full video is above, and a detailed summary is below.
I believe that AI can have the same scale of positive impact on human flourishing as low-cost energy—but only if we properly value it, integrate it, and liberate it.
Properly valuing AI: AI, like low-cost energy, is a new fundamental of human flourishing
What does low-cost energy do, and why is energy a fundamental that is at the root of everything we think of as a good life today? The key is what I call “machine labor.” Low-cost energy—machine calories—enables us to use machine labor to dramatically amplify and expand our physical abilities.
Low-cost energy amplifies and expands our physical abilities
Low-cost, in the end, means that it does not take a great deal of human time to make something happen. Our ultimate cost is our time. If people can have a great deal of energy without spending much time to produce it, then they gain an army of machine servants that can do all sorts of things they could never do before.
Machine labor amplifies our physical abilities. My favorite example is a modern combine harvester. One person using a combine harvester can reap and thresh as much wheat as it once took roughly 1,000 people to handle. That is why only a small share of the population now needs to work in agriculture. Low-cost energy turns ordinary men and women into supermen and superwomen.
Machine labor also expands our physical abilities. It powers machines that can do things no number of humans can do. If a child is born prematurely, no number of people can get together and function as a human incubator. Even a king 500 years ago could not fly across the country, no matter how many people he could assemble. Machines can do things we simply cannot do.
Low-cost energy amplifying and expanding our physical abilities has made the modern human environment unrecognizably better. Procuring food and water now takes a tiny fraction of our time compared with what it required for most of history. We are dramatically safer from climate, predators, and disease. We have running water, easy travel, modern medicine, and instant communication. And we have far more opportunity because so much of our time has been freed up from basic survival.
AI amplifies and expands our mental abilities
If low-cost energy has dramatically amplified and expanded our physical abilities, then AI can dramatically amplify and expand our mental abilities.
Just as the masses once lacked the physical abundance we now take for granted, most people today lack the mental abundance that the wealthiest take for granted. Wealthy people have access to forms of teaching, advising, and assisting that others do not. AI has the potential to make those benefits available to billions.
Teaching: The best teachers do not just know a subject. They know the specific student. They understand that student’s context, starting point, confusions, and gaps in knowledge. That is why great teachers and tutors are so valuable. AI is getting to the point where it can provide a deeply customized explanation about virtually any topic to an individual learner. This is a world-changing improvement in teaching.
Advising: Much of the value of an advisor comes from applying expert knowledge—whether in medicine, law, psychology, etc.—to a specific situation. AI is increasingly able to provide useful, tailored advice, at least as a supplement, at a cost that is vastly lower than what human experts charge. That means billions of people will gain access to forms of guidance that today often cost hundreds or thousands of dollars an hour.
Assisting: In work and in life, people benefit enormously from having more junior helpers who free them up for higher-level tasks. That is true in programming, research, writing, law, and many other fields. AI is already acting as an invaluable assistant for many people—saving their time, increasing their productivity, and allowing them to focus more of their effort on higher-value work.
AI will not only create more mental abundance, it will also compound physical abundance, because physical abundance ultimately depends on mental labor. The internal combustion engine was created by minds. Battery improvements come from minds. Better power systems come from minds. If AI amplifies and expands human mental abilities, it will accelerate improvements throughout the physical world.
The human flourishing potential of AI matters enormously because most of the world is still poor by our standards. Even after adjusting for purchasing power, the average world income is still extremely low relative to what many in the developed world consider normal. The world is not running out of things to do. The world is full of unmet needs. We should be deeply excited about a technology that can create such extraordinary mental and physical abundance.
Just as we cannot think clearly about fossil fuels without recognizing that the modern world depends on them, we cannot think clearly about AI without recognizing the scale of value it can provide. One day people may look back and find it hard to imagine a world without customized geniuses teaching them, advising them, and assisting them. But that will only happen if we first learn to see AI’s value clearly.
Properly integrating AI: AI will reach its potential only if we use it according to its strengths and limitations.
Once we properly value AI, we must learn to properly integrate it into our work. This is one of the key practical challenges with getting the most out of AI right now.
The Industrial Revolution did not work by creating machines that simply replicated human beings. Machines had different strengths and weaknesses, and people had to learn how to use them intelligently. The same is true of AI.
AI has extraordinary capabilities, but it also has real limitations. It can simulate intelligence in powerful ways, but it does not have an independent connection to reality. That means human beings still need to validate many of its outputs.
AI is already changing the nature of work. In many cases, AI can do a great deal of front-end research or drafting, but humans still need to check the results. That changes the division of labor. It does not eliminate the need for judgment. It changes where judgment is most needed.
A common mistake is to overuse AI by treating it exactly like a competent human employee. Another common mistake is to underuse AI because it makes some obvious errors. Both approaches fail to recognize that AI is a powerful force with unique strengths and limitations, and it needs to be integrated accordingly.
Properly liberating AI: AI will fall short of its promise if it is trapped by bad policy.
Once we properly value and integrate AI, we also need to properly liberate it. Freedom is a requirement of progress in every field, and AI is no exception.
Today, AI faces serious political threats, such as opposition to data centers, opposition to industrial development, and restrictions on permitting. If we continue to impose severe industrial restrictions, we will fall far short of AI’s potential. This matters first because life will be far worse than it could have been with AI freedom. And it also matters because technological underperformance can become a security issue in a hostile world.
Like low-cost energy, AI is already under irrational and anti-human cultural and political attack.
Much of today’s popular opposition to AI is reminiscent of the popular opposition to fossil fuels that we’ve heard for decades.
In particular, critics of AI often fail to acknowledge the enormous potential benefits of AI and instead focus only on the negatives. One version of this is the abuse-use fallacy: the idea that because a technology can be abused, it therefore should not be used.
AI can certainly be abused. But that does not change the fact that it can also become one of the greatest tools for informing, teaching, advising, and assisting people in human history.
AI needs to be championed from a confident human flourishing perspective.
It is not enough to defend AI as a necessary evil. The same was and is true of fossil fuels. For years, many defenders of fossil fuels accepted the premise that fossil fuels were dirty, regrettable, or dangerous, while arguing only that society still needed them for now. But fossil fuels are not a necessary evil—they are a profound good that have made human life radically better. The same is true of AI.
AI is not an unfortunate necessity, but a profound potential good. It is a new fundamental technology with the power to make people more capable, more productive, and more knowledgeable.
We are at a crucial point in the development of AI. We have a new fundamental technology, but many people are thinking about it through the wrong philosophical framework. If we think about AI through the right framework—on the standard of human flourishing—we can value it properly, use it properly, and fight for the policies needed to unlock its full potential.
That’s what I’ll be doing, and I hope many of you will join me.
Questions about this article? Ask AlexAI, my chatbot for energy and climate answers:
Popular links
EnergyTalkingPoints.com: Hundreds of concise, powerful, well-referenced talking points on energy, environmental, and climate issues.
My new book Fossil Future: Why Global Human Flourishing Requires More Oil, Coal, and Natural Gas—Not Less.
“Energy Talking Points by Alex Epstein” is my free Substack newsletter designed to give as many people as possible access to concise, powerful, well-referenced talking points on the latest energy, environmental, and climate issues from a pro-human, pro-energy perspective.
