The United States Must Reject Government Control of Artificial Intelligence
Every major technological revolution—but especially information and communications technology (ICT) revolutions—eventually feature calls for far-reaching political controls. The argument is always the same: powerful technology must be centrally managed, or potentially even owned by the government, in the name of safety, security, job protection, wealth-sharing, etc.
Artificial intelligence (AI) has become the latest example of this trend. In a remarkably short time, the Overton Window in AI policy has shifted multiple times. Examples include extreme calls for global development bans, and even potential nationalization following the debut of ChatGPT, and then again more recently following Anthropic’s warning that it viewed its new Claude “Mythos” model as too powerful for immediate public release.
After a brief period of policy moderation that featured efforts by the Trump administration and other lawmakers to encourage wider AI development, extreme proposals are back on the menu. Sen. Bernie Sanders (I-Vt.) argued recently that the government should own a 50 percent stake in large AI companies. Recent reports also state that some senior Trump administration officials have held preliminary discussions with major artificial intelligence companies like OpenAI and Anthropic about potential “partnerships” with the federal government to acquire shares in their firms or share profits in some fashion.
No matter what form nationalization takes—whether “hard” (outright government ownership) or “soft” (equity stakes or wealth-sharing “partnerships”)—these proposals share the dangerous common premise that AI is too important to be left to competitive markets and decentralized experimentation. They all entail government controlling algorithmic and computational technologies through top-down plans and pressure.
Whether this is being proposed by progressive liberals, conservative populists, national security hawks, or even AI companies themselves, all varieties of AI socialism or state corporatism must be rejected. The history of government ownership or centralized control of important technologies, and especially ICTs, illustrates the serious dangers for competition, innovation, national competitiveness, speech, and freedom more generally. There are also far more sensible ways to address concerns about AI systems—such as broad-based taxes or generally applicable regulations—without resorting to radical solutions like government-ownership.
Foreclosing Innovation and Competition
The most obvious danger of AI socialism is that it will undermine competition and innovation. Markets are a discovery process featuring an ongoing interplay of ideas, business models, technologies, and firms that compete to bring new products forward. Capitalist systems encourage repeated acts of entrepreneurialism and “creative destruction.” This constant churn helps determine which ideas work and which do not. Growth and innovation flourish in such environments and a greater diversity of economic and cultural output results.
AI nationalization would substitute political decision-making for this process and stagnate markets. Government co-owned AI systems would inevitably require officials to determine which firms deserve support, which technical solutions get developed, and which applications are socially acceptable. These decisions would no longer emerge from decentralized interactions but through administrative proceedings or influence. Government officials would be picking winners and losers.
This would represent a stark reversal of the highly successful policy model that helped America’s technology sector become a global powerhouse in computing, communications, software, internet services, and now algorithmic technologies. The central lesson of the internet revolution is that innovation flourishes where entrepreneurs and investors are free to experiment, iterate, fail, and try again. This culture of permissionless innovation helped generate continuous waves of new ideas and firms. Some of today’s leading AI firms and models were completely unknown to the public just a few years ago.
AI nationalization would end this by requiring extensive review procedures, licensing and compliance mandates, and other centralized oversight structures. Bottom-up discovery becomes far less likely in that environment because bureaucracies excel at preserving existing structures and procedures. Disruptive change and the benefits associated with competitive rivalry become harder, and eventually regulated monopoly sets in.
Centralized Control Breeds Politicization and Cronyism
This points to the important second-order politicization effects that would be associated with efforts to nationalize AI, such as regulatory capture and cronyism. Rather than thousands of entrepreneurs pursuing diverse innovations and new business models openly in markets, AI would be shaped by a smaller group of regulators, bureaucrats, and politically connected players. The concentration of decision-making power would create perverse incentives for firms and special interests to focus more on pleasing politicians to stay in their good graces and retain the benefits of continued protection or procurement contracts.
Various OECD reports on state-owned enterprises have identified how they “suffer from undue hands-on and politically motivated ownership interference, leading to unclear lines of responsibility, a lack of accountability and integrity and efficiency losses in their corporate operations,” and that they “remain vulnerable to being used as conduits for political finance, patronage, and personal or related-party enrichment.”
Even lesser forms of government control breed similar problems. In his 1971 magnum opus on The Economics of Regulation, economist Alfred Kahn noted of public utility regulation that:
When a commission is responsible for the performance of an industry, it is under never completely escapable pressure to protect the health of the companies it regulates, to assure a desirable performance by relying on those monopolistic chosen instruments and its own controls rather than on the unplanned and unplannable forces of competition. […] Responsible for the continued provision and improvement of service, [the regulatory commission] comes increasingly and understandably to identify the interest of the public with that of the existing companies on whom it must rely to deliver goods.”
In the late 1970s, President Jimmy Carter tapped Kahn to help address such problems at the Civil Aeronautics Board (CAB), which regulated airline routes and prices. Kahn, a lifelong Democrat, concluded that regulatory capture was so entrenched at the CAB that abolishing the agency was the only effective way to advance consumer welfare. He then worked with fellow Democrats in the Carter administration and Congress to dismantle the agency and eliminate its protectionist, anti-consumer policies.
The same sort of capture was on display over the past century within American communications and media technology markets, where regulators and firms centralized control over information systems through government ownership of the airwaves, licensing regulation, price and entry controls, and content controls. Government-endorsed protectionist schemes created regional and national monopolies protected from the rigors of competition. Markets stagnated, and consumers lost new options.
Extending that model to AI would risk recreating the same problems on a much larger scale and would undermine what has thus far been one of the most vibrantly competitive and innovative technology sectors in history. AI firms would instead engage in a race to curry favor with government officials, build a moat around their products, and then pull the drawbridge up behind them once their castles enjoyed government protection.
Undermining Global Competitiveness
AI nationalization would also weaken America’s geopolitical standing. The United States is engaged in a global race for leadership in advanced computation. Nations around the world recognize that AI capabilities will shape their economic and military destiny. “AI competitiveness matters because AI capabilities and capacity affect nations’ abilities to pursue strategic goals, such as economic growth or national security,” notes a new U.S. Government Accountability Office report.
Government-directed AI systems would erode America’s core strengths over the rest of the world by undermining private-led creativity, competition, and investment. The United States became a technological leader because it developed policies that encouraged robust private-sector entrepreneurialism and openness to new ideas and change. Private AI capital expenditures has been “a massive private sector stimulus program,” boosting domestic economic growth, and strengthening American global technology leadership.
The best response to foreign technological competition is not to adopt the centralized approach favored by many authoritarian regimes. It is greater dynamism to ensure the stock of vibrantly innovative new firms and products is constantly restocked. America cannot rest on yesterday’s victories if it hopes to remain on the cutting-edge of the technology frontier.
Threats to Freedom of Speech and Expression
The final danger of AI nationalization involves the freedom of the public to use different AI models or applications that are free of government control or influence.
AI has the potential not only to be the most important general-purpose technology of our era, but also the most important information technology. It could become a technology of freedom that empowers the public and gives citizens greater ways to create, access, organize, and disseminate knowledge. Alternatively, AI could have a darker future as a technology of control, where governments seek to limit options or even “weaponize” algorithmic capabilities against citizens.
The history of governments’ control over ICTs is littered with examples of censorship, surveillance, propaganda, and jawboning. This is why government control over AI—especially nationalization proposals—should concern anyone who values freedom of thought and expression. If the government demands “partnerships” or takes stakes in major AI companies, control points will multiply and opportunities for abuse will expand. Decisions about model design, content policies, information access, and acceptable outputs all become political decisions. Unfavored groups or ideas will be more easily penalized.
A decentralized AI ecosystem allows for multiple models, competing perspectives, and diverse sources of knowledge that provide important safeguards against concentrated power. AI nationalization would move society in precisely the opposite direction and create an information marketplace that looks more like something in line with the current Chinese Communist Party vision for centralized control of truth instead of the American tradition of openness and the free flow of ideas and expression.
Greg Lukianoff, President and CEO of the Foundation for Individual Rights and Expression, has warned that “a government monopoly on advanced AI,” or even a government-empowered oligopoly that concentrates informational power in fewer hands, would create “the framework not only to censor but also to dominate and distort the production of knowledge itself.”
The template for this already exists. “Nationalization of AI will accelerate the corporate-government fusion we’re already sliding toward,” warns David Sacks, President Trump’s AI czar. The Trump administration has obtained direct equity stakes, or “golden shares,” in a dozen companies over the past year, including U.S. Steel and Intel. Government stakes in firms creates added dangers when speech-related outputs are involved. Sacks cautions his fellow conservatives to be “concerned about Central Government AI — a system with even more totalistic power over information, decision-making, and human behavior” and to “think more carefully about how regulations they are flirting with now.”
It is wise advice. The Trump administration has already used its special stake in the Nippon Steel–U.S. Steel partnership to prevent the company from closing an unprofitable plant in Illinois. If politicians have such influence over AI companies through “partnerships,” similar pressure could be applied by government agents to control content decisions.
This is a continuing problem in traditional ICT markets, where policymakers still today look to use control mechanisms to both directly and indirectly harass opposing viewpoints. For example, Federal Communications Commission (FCC) Chairman Brendan Carr has abused the agency’s amorphous powers to threaten the revocation of broadcast licenses based upon his anger over their coverage of the Iran war or the anti-Trump comments of late-night talk show hosts. Carr has even bragged about getting “wins” over the media on behalf of Trump and claiming he has the right to police “fake news” or “news distortion.” It makes no difference that this behavior comes wrapped in the cloak of “the public interest” when lawmakers or regulators are using political mechanisms to determine what they think that means.
If government officials come to have similar sway over AI firms, it is likely such intimidation tactics will be used to influence the output of their models and directly or indirectly control speech in the process.
Conclusion: Keep AI a Technology of Freedom
Nobel Prize-winning economist Friedrich Hayek famously dedicated The Road to Serfdom, his 1944 classic critique of central planning, “To the socialists of all parties.” Hayek warned all those seduced by the allure of top-down planning that such control efforts would inevitably devolve into repression, stagnation, and a loss of freedom broadly. Echoing Hayek, the Wall Street Journal editorial board argued that AI nationalization represents “the road to AI state socialism.”
Centralized government control of AI would undermine the very forces that made America the world’s technological leader: competition, entrepreneurial experimentation, dynamism, and freedom of thought. AI socialism must be rejected to preserve our freedoms and ensure America stays on the cutting-edge of the most important technological revolution of modern times.