The debate on AI in the United States has gone in an interesting direction – one that’s nearly radical in its simplicity: what if the American people owned something from the AI boom? President Trump has been thinking about this very thing, proposing that the top AI firms should “return to society” of which they are in many ways a product. The idea has generated an intriguing discussion among tech companies, legal experts and policy makers on the exact nature of how such a system could be implemented as well as whether it could influence the relationship between government, industry and citizens in the digital era.
The concerns are the reason for this conversation and are not difficult to understand. Over the next few decades, artificial intelligence is expected to create a trillions of dollars in wealth, but there is a fear that this wealth will fall into the hands of a small group of tech execs, VC investors, and early investors. Meanwhile the average American worker could be at risk of being replaced by the very technology that is creating still more wealth for others. It is not a new phenomenon, economists have bemoaned the phenomenon of inequality caused by artificial intelligence for years, and the filing of IPOs by OpenAI and Anthropic, with OpenAI planning a listing valuing up to $1 trillion, has only added to those fears. In a company with such stratospheric valuations, the question of who benefits cannot be ignored.
One of the most sweeping plans is that of Vermont independent Sen. Bernie Sanders. Sanders has proposed that big tech companies be subjected to a 50 percent stake in their companies and board seats, ensuring that the American people would have a stake in and a say over how those companies are run. The initiative is ambitious and has been welcomed and challenged by people from all sides of the political divides. The senator’s point is not only about the economic benefits of AI but also about how the American people can take part in and hold accountable what’s wrong.
The Sanders proposal is based on a previous academic paper by two law professors suggesting a tax system in which bilateral deals could be made with the tax payment in shares instead of cash. Such a method would be a way to shift equity to the government, albeit without any public spending, and would be akin to a royalty on the value of an enterprise for its use in the American economy. That would not strip the government of its control of these companies, which could make it easier for investors and corporate heads to take the plunge, as they tend to be wary of government interference in business operations. One important element of this model is that it is elegant—it allows for public benefit without the direct ownership of governments that limits innovation and prevents private investment.
A second route that has been followed is based on a recent arrangement made by Intel where the government based in Silicon Valley acquired a 10% stake in Intel with billions of dollars invested in a capacity expansion program at home. This is not the case in this model, and maybe this would be a good fit for the AI industry as it has an appetite for capital. The tech industry has been spending huge amounts of money on AI infrastructure – including big data centres, dedicated chips and cooling systems. Government funding, in particular, can play an important component in that, and it can be stable and large enough, and it can be shared with the public so that the public shares in the rewards. The key benefit of such an approach is that it has been applied in practice and can be implemented in the semiconductor industry, which has many common issues regarding capital intensity and national competitiveness.
The discussion also has incorporated the concept of having U.S. government personnel on company boards to provide the public with a voice in the development and deployment of these powerful technologies. It is not a mere question of monetary profit, but one of governance and oversight, the unease over the societal repercussions of AI systems increasingly making decisions that impact human lives significantly. Board representation would provide a mechanism for accountability which purely financial stakes do not, and in doing so ensure public concerns about bias, privacy and safety are heard at the highest levels of corporate decision-making.
For now, however, the big AI companies – such as Anthropic, Google, and OpenAI – have stayed mum on these proposals. Their lack of comment may be understandable given the sensitivity of the issue and the complexity of the issues. Businesses that are operating in the midst of rapid expansion and heavy competition are faced with unprecedented challenges and the potential for government involvement, including questions of control, valuation, and even the nature of the business itself. Meanwhile, the tech industry has been pushing back against government concerns on matters that impact the nation as a whole, ranging from cybersecurity to semiconductors, which could open up opportunities for negotiation and innovative solutions.
The concept that the American people will have a stake in the success of AI companies is not socialism but isn’t even capitalism, per se. It recognizes the American people have invested heavily in research and education to support the development of AI through their tax dollars and public institutions. It also acknowledges that the advantages of technological advances are not to be confined to a small group, particularly when the risks and disruptions are widespread. However, these proposals are subject to relevant concerns of government overreach and the potential for chilling effect on innovation that may result from too much regulation or ownership. Not only that, but it would be essential to determine how these stakes would be valued, managed, and eventually transformed into benefits to the public.
