SIrricon Valley continues to ride the wave of Wall Street’s never-ending frenzy. Investment in artificial intelligence continues to exceed expectations, with Bloomberg reporting that financial firms plan to pump about $700 billion into the industry this year alone. AI stocks currently account for about one-third of the total stock market and 45% of the S&P 500’s market capitalization.
Of course, the extraordinary investment in AI is driven by the promise that this technology will eliminate the need for large numbers of workers. Leaders in the technology industry aren’t shy about saying a lot. “It is not clear where these people will go or what they will do. We are concerned that they may become unemployed or form an ‘underclass’ with very low wages,” says Dario Amodei, head of Antropic.
If that’s not alarming enough, the explosion in AI investment signals perhaps the largest speculative bubble in history. If it ruptures, which is a very real possibility, the entire world economy will collapse. In any case, the AI boom and bubble represents a tremendous social and economic responsibility, one that governments will ultimately have to consider. But how?
In America, the standard policy response to any kind of economic crisis is to leave workers to fend for themselves and bail out failing companies. President Barack Obama’s automobile industry policy is a prime example. In one sense, this reasoning was valid, albeit terribly imperfect. The fall of Detroit would have meant destitution for hundreds of thousands of autoworkers and millions of other workers dependent on the industry. But once the checks were cashed and the problems resolved, the government and the governed failed to reap significant returns on their public investments, despite at one point owning 61 percent of General Motors’ stock. No dividends or profit distributions were made. A few years later, General Motors launched a massive stock buyback program worth $10 billion in 2023 and $6 billion in 2024. All benefits were increased to maximum levels, while workers were required to make “common sacrifices” on the job.
This doesn’t work with AI. Not only are bailouts economically insufficient to rescue severely overcapitalized companies, there is also no political or moral legitimacy to bail them out. In fact, it’s not clear whether AI successes or failures will result in even more jobs being lost.
Addressing AI therefore requires starting from a different, bolder premise. Some people want to simply halt the progress of AI. As someone who blames smartphones for many of our social pathologies, I have a lot of sympathy for these neo-Luddites. But while screaming “No!” may be satisfying, it doesn’t do much. For every small town that puts up a barrier to building a data center, there will be a dozen investment-hungry towns begging to cut the ribbon. And even if we somehow put the heads of Anthropic, OpenAI, and Gemini in jail and made AI development illegal in the United States, new companies would soon take their place overseas. The economic incentives are too irresistible. The cat really came out of the bag.
On the other side of the neo-Luddites are technology-friendly progressive reformers who adopt the logic of meteorologists. They argue that you can’t stop or change the course of an AI, just as you can’t stop or change the course of a hurricane. Our only recourse is to build metaphorical levees to slow the rising flood of unemployment claims. Mitigating the devastating effects of mass unemployment through cash transfers. And clean up the mess later. Of course, there’s nothing wrong with trying to help evacuated people through AI automation, just like we did with people evacuated by hurricanes. But proposals for reform along these lines acknowledge the great destructive lie: that technological progress and its effects are inevitable and uncontrollable—a force of nature. But we can’t stop hurricanes, for example, because we didn’t create them. Engineers don’t design systems. People don’t own hurricane companies.
The development of AI is not like a hurricane, it is not natural. A side effect of technological progress is the result of decisions, decisions made by the AI ownership class. These men are in destructive competition with each other for unimaginable wealth, controlling the course of things and thereby our collective economic destiny. Advocating for reactive reforms such as universal basic income and job retraining obscures this fundamental fact. And it lends credence to the Silicon Valley myth that, whatever the disruptive effects of a particular technology, this is an unfortunate but necessary price to pay for progress. Worse, this inherently defensive posture will do little to calm the coming economic storm.
This doesn’t mean there’s nothing we can do. Once we understand that technological progress cannot be completely suppressed and that there is no single predetermined path for such progress, we can begin to think beyond pauses and moderation. Instead, we should look to use this tool for social purposes. So what is a better foundation on which to build more proactive policies?Public ownership.
That may sound extreme, but it’s not that unthinkable. Consider that if the AI bubble bursts or the economy collapses due to widespread wage depreciation due to forced displacement of workers, public calls for some form of social relief will become even louder and big ideas more appealing.
More importantly, we have a moral claim to property rights. As a society, we helped make AI a reality. Government investment is not only helping to revitalize and sustain the technology sector; AI’s success is uniquely built on the uploading of our collective work. This year, Anthropic ripped off the spines of millions of print books (some long out of print), scanned them, and threw them away to “train” a large-scale language model. That’s a scandal. Not just bibliophiles. Machine learning tools are currently “listening” to millions of hours of human-generated music to generate new fully manufactured AI clones. The same goes for the generated videos, which can imitate the style of a particular director or resemble a particular actor after siphoning off the entire corpus of digitized videos. A collective of countless writers, thinkers, crafters, artists, photographers, filmmakers, craftsmen, farmers, gardeners, designers, architects, etc. are uploaded into its “brain.” These millions of human “teachers” will forever be discredited and will certainly not be paid for their services. Their contributions, however modest, have yielded dizzying returns for the likes of Dario Amodei, Sam Altman, Elon Musk and Mark Zuckerberg.
Not only do we have a moral basis for public ownership, but we also have a clearly patriotic basis because of our collective contribution to public ownership. In short, public ownership of artificial intelligence is in our national interest. Reports have portrayed the so-called AI race as a geopolitical competition between the United States and China to see who can invest more, learn faster, and build the best models. But this is also a myth. With some exceptions in China, the AI race is not between nations but between private companies. Therefore, the spoils are not returned to the citizens or the popular class, but only to the oligarchy. This is especially true in the United States. In the United States, the nation invests directly in technology, ensuring a favorable business environment for AI expansion through a biased regulatory environment and massive tax incentives. Meanwhile, the only advantage of the popular classes seems to be the privilege of paying a small amount of money to slowly train your brain with slops of vertical videos.
There lies a great contradiction between, on the one hand, the rights and interests of the popular class expressed through the democratic nation-state, and, on the other hand, the free property rights on which that nation is ostensibly constructed. Despite the direct investment in infrastructure that made these technological developments possible, and despite the expropriation of our collective intellectual work by machine learning, we as citizens receive nothing in return. At the same time, the high-tech oligarchy that sits above us has no lasting allegiance to the democracies to which it owes its success. Evidence of these divergent interests goes beyond even partisan lines, as both Donald Trump and Bernie Sanders have called for the nationalization of AI companies. On the other hand, the very fact that Mr. Amodei feels no obligation to comply with President Trump’s demands for Anthropic’s backend confirms that the tech elite has power beyond the state.
The easiest and fairest way to solve this conundrum is to reconcile these conflicting interests by establishing a general direction for AI companies. Not surprisingly, Mr. Sanders himself introduced legislation that would require the largest AI companies to transfer 50% of their equity to public funds. America’s AI sovereign wealth fund would function much like the funds set up by Norway and Alaska for their oil and gas reserves. Because the value of natural resources such as oil and gas does not depend on the ingenuity of any particular company or individual, their value itself is naturally considered to be common property. The same goes for AI. The difference, however, is that the value of technology comes from human collaboration. Shouldn’t it be collectively governed and its profits distributed that way?
Despite the claims of critics, this is not a handout proposal, quite the contrary. Public ownership of AI represents an effort to reclaim socially generated wealth for social purposes instead of philanthropy. In the process, we may be able to right the economic wrongs of the general mismanagement of the knowledge economy. There is a yawning wage gap between the non-university-educated working class and the university-educated working class that has broadly benefited from the technologicalization of the economy. Just as free trade and the China Shock eroded wages for industrial workers, we now face a new crisis as advances in AI threaten the wage premium for college-educated workers. In this way, AI, the apotheosis of the knowledge economy, could offer a way out of the complete domination of domestic politics by high finance and Big Tech.
As the explosion of public debt threatens domestic social policy, democratic leaders must begin to reimagine ways to regenerate societies that provide states and their people with independent means to achieve social objectives. If the bubble bursts this time, the response must be aimed at correcting, not exacerbating, the power balance between capital and labor. The wealth generated by AI-related developments can be reinvested in social renewal programs, such as investing in infrastructure to update the built environment, funding for primary and secondary education, new energy production, advanced manufacturing, social services, etc.
Public ownership of AI could therefore be not only a moral way to address the upcoming employment and social crises, but also a transformative attempt to restructure the economy for the common good. In that respect, the oligarchy is right to be afraid.
