Three paths forward, each challenging what we value most.
I suspect artificial intelligence may eventually accomplish what generations of socialists have struggled to achieve. And the irony is that capitalism may ultimately be responsible for making it happen. And I say this as someone whose political instincts lean libertarian.
I have never been particularly enthusiastic about government-supplemented incomes, wealth redistribution, or expanding the state’s role in economic life. Yet the more I consider the potential consequences of artificial intelligence, the more I find myself confronting this uncomfortable possibility.
The economic arrangements I have long defended, particularly free markets and the principle of individual economic independence, may become difficult to sustain, not because their underlying principles have been disproven, but because the technology we are developing could fundamentally change the conditions upon which they depend.
What troubles me is that AI itself does not decide who benefits from its capabilities or who bears the costs. People do, through choices shaped by ambition, self-interest, competition, and sometimes greed — well, mostly greed. We may build machines capable of extraordinary intellectual achievements without becoming any wiser about the desires that govern their use.
We appear to be heading toward a future in which artificial intelligence allows companies to produce much of what we need with a fraction of today’s workforce. Previous technological revolutions displaced workers but also created new occupations, and if you’re feeling optimistic, AI may do the same.
But that optimism rests on the assumption that AI will follow a familiar historical pattern. Unlike previous technologies that automated particular forms of work, AI can perform an expanding range of intellectual tasks across industries, including some of the very capabilities people have historically relied upon to adapt to technological change.
What happens to an economic system built around the relationship between work and income when fewer people are needed to produce the goods and services society consumes? And what happens to our philosophical convictions when the conditions that once justified them fundamentally change?
As I see it, such a future points toward three broad directions, each requiring us to sacrifice one value in pursuit of another. Of course, some would argue that compromises are possible between the scenarios I’m about to present. Yet I question whether such compromises would offer more than temporary relief from the underlying tensions.
The First Path: Continue as We Are
The first, and perhaps most tempting, path forward is to take the path of least resistance and carry on as we are, allowing businesses to use AI as extensively as they choose while leaving our employment-based economy largely intact. But for such an arrangement to hold, we would have to trust businesses not to replace too much human labor, even when doing so would save money or improve performance. And to that, I say, good luck.
Having spent much of my career in business, I recognize that cutting costs is not always a matter of greed. Some companies may pursue AI to maximize profits, while others may adopt it simply to survive.
Consider what happens when one company discovers it can produce the same goods or services with fewer employees. Its competitors face pressure to achieve similar efficiencies or risk falling behind, even if they would otherwise prefer to retain their workers. As more businesses make these individually rational decisions, the cumulative effect could gradually undermine the relationship between employment and income on which our economy depends.
Governments could attempt to prevent this through regulations or financial penalties for replacing human workers with AI. But how effective would such measures be? Businesses have spent generations minimizing taxes, relocating operations, lobbying lawmakers, and finding ways around regulations.
Companies could move automated operations overseas, restructure positions to avoid penalties, or gradually transfer responsibilities to AI without formally eliminating jobs. Governments could respond by tightening restrictions, but we could spend decades closing clever loopholes as new ones emerge.
The larger problem is that workers are also customers. Their wages allow them to purchase the goods and services businesses produce. If AI reduces the need for human labor across enough industries, we could find ourselves with extraordinary productive capacity but fewer people earning the income necessary to participate in the economy.
No individual company would necessarily be responsible for this outcome, nor would hiring workers it no longer needs offer a practical solution. What makes economic sense for businesses individually could ultimately undermine the economic conditions on which they collectively depend.
Continuing as we are might preserve the freedom of businesses to pursue efficiency, but at the cost of the economic security and independence that employment has traditionally provided. That is a considerable sacrifice for a system that places such importance on earning one’s own living.
The Second Path: Turn Away from AI
Rather than continuing down our current path or relying on targeted government regulations to discourage businesses from replacing workers, we could take a far more drastic approach and severely restrict the development and use of AI altogether. The objective would be to preserve the jobs on which economic independence currently depends.
I understand the appeal of protecting people’s ability to support themselves through their own work, but consider what such a decision could cost us. AI might accelerate cancer research, develop more sustainable ways of producing food and energy, expand access to knowledge, and help solve problems that have resisted us for generations. I find it difficult to justify sacrificing those possibilities simply to preserve the way we currently earn a living.
The difficulty is that there is no obvious way to separate acceptable uses of AI from unacceptable ones. We might welcome a system that advances scientific research while objecting when the same capabilities replace analysts, administrators, or other workers.
Progress in one field can also disrupt employment in another. Someone would have to decide which benefits justify that disruption, which jobs deserve protection, and how much innovation society should forgo. Even if we agreed on those judgments today, new applications would make those boundaries harder to maintain tomorrow.
There is also the question of whether such restrictions could endure. The impulse to invent, explore, and overcome limitations is as much a part of human nature as the pursuit of wealth or status. Nations, businesses, and research institutions would have powerful incentives to continue developing AI, particularly if their competitors were doing so. This path might preserve familiar forms of employment, but it would also require us to deliberately constrain a technology whose greatest benefits we may not yet understand.
The Third Path: Reconsider What Liberty Requires
That leaves the possibility I find most uncomfortable. We continue developing AI, accept that substantial labor displacement may occur, and reconsider how people obtain the resources necessary to live. Some form of guaranteed income, broader ownership of productive technology, or another mechanism may be required.
I recognize that not every form of income support constitutes socialism, and I do not claim to know which approach would work best. But the deeper question is philosophical, not merely political. An economy built around the expectation that people earn their own living may have to confront a future in which opportunities to do so become increasingly scarce through no failure of individual effort.
My instinct is to resist any arrangement that makes people more dependent on government. Yet the freedom to support oneself presupposes a meaningful opportunity to do so. Keynes imagined technological progress relieving humanity of much of its economic burden. We may achieve something resembling that ambition while retaining an economic system that requires most people to sell their labor to survive. We could make work less necessary without making workers any less dependent on having work.
I should make something clear. My willingness to consider greater economic intervention is not an endorsement of socialism. And if you believe that replacing corporate power with government authority somehow eliminates the human appetite for wealth, power, and control, you are fooling yourself. Governments are no more immune to corruption than corporations, regardless of the virtues their political ideologies claim to represent. Changing the institutions that distribute resources does not change the human beings who operate them.
My instincts are closer to subsidiarity, the principle that individuals and smaller communities should address problems within their capacity, with larger institutions stepping in when necessary. That could mean exploring local cooperatives that own AI tools, community land trusts, or regional funds that share the gains from automation rather than assuming any response must come from a central government. I want people to retain as much freedom and responsibility over their lives as possible.
But if the means of earning a living become unavailable to large numbers of people, the problem may exceed what individuals or local communities can solve. Accepting that possibility would not require me to abandon liberty as a value. It would require me to question whether the institutions I have preferred can still protect it.
Of course, new occupations may emerge, and the scale of displacement remains uncertain. This is a thought experiment, not a prediction that human work will disappear. Still, if AI makes labor substantially less necessary, we may face an economic transformation that none of my political instincts would have led me to embrace.
What Are We Willing to Surrender?
The first path asks us to accept the possible erosion of economic security in pursuit of greater efficiency. The second asks us to restrain our impulse to invent and discover. The third asks us to reconsider how much institutional dependence we are willing to accept in order to preserve economic independence. None is comfortable. Perhaps the deeper challenge is recognizing that the political arrangements I associate with liberty are not necessarily identical to liberty itself.
I remain hopeful about what AI could make possible, and deeply uneasy about the society we might build around it. I have no desire to surrender my belief in individual liberty, but I am beginning to question whether the economic arrangements I have long defended will remain capable of protecting it. Perhaps the real test of our convictions is whether we can recognize when the means we have chosen no longer serve the ends we value.
Don’t miss out! Join my email list and receive the latest content.
What are we willing to sacrifice for AI? was originally published in UX Collective on Medium, where people are continuing the conversation by highlighting and responding to this story.
