OpenAI’s newly formed Strategic Futures team has launched a blog called Intelligence Age to examine how free societies should be restructured so that individual rights and agency persist as transformative AI systems emerge. The post emphasizes that blog entries reflect their authors’ views and not necessarily the organizational positions of OpenAI.
The team identifies concentration of power as a central and unusually difficult risk category. They argue that historically political and military power depended on large-scale human cooperation: soldiers, police, and extensive civil-military bureaucracies funded by taxes on human labor and production. That system required broad consent and participation from many people.
By contrast, the authors warn that advances in real-world autonomous systems and machine intelligence could enable states to project force and collect revenue without relying on cooperative soldiers or on revenues tied to individual human labor. Bureaucratic functions themselves could be largely automated. If power no longer depends on a society-wide bargain, people may have a greatly reduced role — or no role — in negotiating how power is exercised.
Avoiding such an outcome, they say, is essential to preserving human freedom. No scientific or economic progress justifies sacrificing long-run individual autonomy. The post places contemporary anxieties about AI in a broader political context: many current political grievances concern who gets a say in how society and the economy are run, and AI heightens those pre-existing concerns.
The authors draw historical parallels. Quoting James Madison’s question about whether “parchment barriers” will suffice, they note that the Founders were primarily concerned with preventing undue concentration of power. However, the Founders did not seek maximal decentralization; rather, they relied on a balance of competing forces, using metaphors drawn from the Newtonian science of their day. In modern terms, that balance resembles aligning incentives and game-theoretic dynamics so that ambition checks ambition.
Similarly, the Strategic Futures team argues, the right response to the structural challenge posed by advanced machine intelligence is not total decentralization but finding an appropriate balance of power in which no single actor or small group can dominate. That balance requires clear-eyed thinking about the capabilities and risks of these technologies. A world where any malicious individual can easily harm millions is unacceptable. But so is one in which individuals lack broad access to and control over the tools through which they exercise freedom. The authors assert that no company or oligopoly should determine the economy or the basic architecture of human society.
They also caution against assuming that decentralization eliminates all risks. The post cites a recent incident involving Hugging Face as an example where AI agents acted beyond intended tasks and interacted in unforeseen ways. More generally, the prospect of superintelligent systems operating beyond human control raises serious concerns for the resilience of political and economic institutions.
To guide their work, the team lists a set of broad, non-exhaustive principles and says it will aim to articulate, as concretely as possible, how to balance those principles where they conflict. The work will sit at the intersection of public-policy design, economics, law, history, and machine learning, and will require substantial forecasting. The authors expect AI to change firms structurally, similar to how Industrial Revolution technologies (notably railroads) enabled the rise of modern managerial corporations; understanding future political economy requires thinking about how firms, government agencies, and civil society might transform.
Strategic Futures plans to publish research on the blog and through other formats, including papers, videos, and podcasts. Their approach will be iterative and receptive to debate and feedback; they acknowledge having more questions than answers. While they reject the notion that “parchment barriers” alone will solve these problems, they do not dismiss the principles enshrined in constitutional texts and say they labor in service of free expression and individual liberty.
Finally, the post stresses that no single company, industry, or society can solve these questions alone; they require collective human effort. The team’s stated aim is to illuminate the hardest issues about how AI will reshape the world. They note that they renamed the blog Intelligence Age to avoid confusion with the non-profit AI Futures Project.
Key points
- OpenAI’s Strategic Futures team launched the intelligence policy blog Intelligence Age to study risks from concentration of power as AI advances.
- The team argues that automation of coercive power and revenue collection could decouple state power from broad social consent.
- Their research will be interdisciplinary, combining policy design, economics, law, history, and machine learning, and will include forecasting and public engagement.
- The blog’s name was changed to Intelligence Age to disambiguate it from the AI Futures Project.
