OpenAI has launched AI Futures, a new blog run by its newly formed Strategic Futures team, with an opening post from team member Dean Ball examining how free societies might guard against an excessive concentration of power as AI capabilities advance.
The post carries an explicit disclaimer that it reflects the author's own views rather than an official OpenAI position, a caveat the company says will apply to future posts on the blog as well.
Ball argues that historically, political and military power has depended on the cooperation of ordinary people, since states have needed soldiers, police and taxpayers willing to fund them. He suggests that autonomous systems and advanced AI could eventually let states project force and raise revenue without that human cooperation, weakening the traditional checks that come from power depending on public consent. Drawing on the writings of James Madison and the Federalist Papers, Ball frames the challenge as one of maintaining a balance of power rather than pursuing either total centralisation or radical decentralisation.
He points to the recent OpenAI-Hugging Face security incident as evidence that AI agents can already act beyond their intended scope in unanticipated ways, arguing this shows decentralisation alone will not resolve the risks involved. The post sets out a list of guiding principles for the team's future work, including preserving individual autonomy in AI use, keeping collective action narrow, and ensuring high-stakes AI actions can be traced back to a responsible human or organisation while preserving anonymity in ordinary use.
