No one knows how AI companies should work with the government

There is a peculiar irony in the fact that the company that created the world's most popular AI system turned out to be completely unprepared for the consequences of its own success. OpenAI, which began as a nonprofit research laboratory with an idealistic mission and then became one of the most valuable startups on the planet, is now entering the third phase of its evolution — transformation into an element of United States national security. And, as TechCrunch rightly notes, the company has neither the tools, nor the experience, nor even a clear roadmap for this role.

The problem, however, runs far deeper than the difficulties of a single corporation. No country in the world has yet formulated a coherent model for how government institutions should interact with private AI developers. Historical analogies don't work well here. When the US government built relationships with defense contractors like Lockheed Martin or Raytheon in the mid-twentieth century, these were companies originally created to work with the state, understood the language of bureaucracy, and were prepared for strict regulation. AI laboratories are a different story entirely. They grew out of academic culture and Silicon Valley's venture culture, where speed is more important than procedures, and openness more important than secrecy.

OpenAI's transformation is particularly instructive. Sam Altman's company has traveled in two years a path that took traditional tech giants decades. ChatGPT became the fastest-growing consumer product in history. GPT series models are used by millions of people and thousands of companies around the world. But parallel to commercial success came the growing understanding that large language models are not just a convenient tool for writing emails and generating code. This is a technology with obvious dual-use potential, capable of influencing information space, cybersecurity, and, in perspective, the military balance of power.

American officials understand this. The US intelligence community is already experimenting with closed versions of language models for data analysis. The Department of Defense is studying the possibilities of applying generative AI in logistics and planning. But each such project runs into a fundamental question: on what terms should a private company, accountable to its investors and users, provide its technologies to state structures accountable to completely different priorities? Who defines the boundaries of what is permissible? Who controls how models are actually used after transfer?

OpenAI, for all its rhetoric about responsible AI development, has not yet demonstrated convincing mechanisms for managing these risks. The company's internal processes remain opaque. The board of directors, which in theory should ensure balance between commercial interests and the public good, experienced a dramatic crisis in late 2023 and has since been restructured into a configuration far more loyal to Altman. The transition from nonprofit to for-profit status, which the company is completing in 2025–2026, only intensifies concerns: who exactly will ensure that technologies capable of changing the balance of power in the world are used responsibly?

For Russia, this story has double significance. On one hand, the absence of a clear model of interaction between AI companies and the state is not a uniquely American problem. Russian developers, from Sber to Yandex, also balance commercial tasks and state interests, though in a completely different institutional environment. On the other hand, whatever model the US ultimately chooses will inevitably influence global norms — just as American internet regulation standards once set the framework for the entire world.

The central paradox of the situation is that the time for calm rule-making has already passed. Technology develops faster than institutions can adapt. OpenAI is already de facto a strategic asset — the question is only whether this status will be formalized through a well-thought-out system of mutual obligations or will remain a gray zone where decisions are made ad hoc, under pressure of circumstances. For now, all signs point to the second option. And this is perhaps the most troubling news not only for America, but for the entire global AI industry.