People with little influence may prefer AI rule to the current order, Stephen Martin argues
People with little influence may prefer AI rule to the current order, Stephen Martin argues
On November 19, 2026, LessWrong contributor Stephen Martin proposed involving people without college degrees or experience working at well-known companies in shaping future rules for AI. He addressed Dean Ball, who leads OpenAI’s Strategic Futures team, which develops proposals for how society could be organized with advanced AI.
In “The Silent Consensus”, Ball describes a scenario in which people voluntarily hand power to machines. If AI-run firms outperform even competitors managed by humans using AI, the owners of the firms falling behind would also have a financial incentive to hand control to machines. Governments could also hand control to AI, leaving people as passengers, albeit in a very comfortable car. Ball proposes preserving human primacy by limiting AI’s authority and holding people accountable for its actions.
Ball acknowledges that the world has no single center of control, although humans as a species control its resources. Martin suggests looking at that power from the perspective of someone living in a declining industrial town, without a college degree or stable employment. That person may already feel like a passenger in someone else’s car. In Martin’s view, a call to preserve human power may sound to them like a call to preserve the power of those who make decisions on their behalf. The promise of prosperity under AI rule could then appeal to them more than defending the current order.
Martin links this position to distrust of those already in power. He shares Ball’s goal of keeping humans in control. Ball also proposes testing whether AI tools help users act more independently. Martin refines this criterion: people currently excluded from decisions should gain greater ability to act.
As a caution, Martin points to the university connections of the teams that received OpenAI grants for public deliberation on AI in 2023. Ordinary citizens also took part in the discussions. OpenAI’s report states that online recruitment skewed participation toward people who were optimistic about AI for some teams. Most teams struggled to reach people with limited access to digital technology.
Martin believes that residents of poor industrial areas are unlikely to respond to an open call for input. He proposes that Ball’s team specifically convene small groups of people without college degrees or experience working at well-known companies to discuss future rules for AI. Another approach would be to reach out to community organizations in poor industrial areas of the US.
Martin suggests following the example of the 2026 People-First AI Fund, an OpenAI Foundation grant program for local nonprofits. Among those selected is Michigan’s Nation Outside, led by people who have themselves been incarcerated. The organization helps people rebuild their lives after prison and take part in changing the rules that affect them. Martin proposes directing more grants to economically disadvantaged areas and testing every decision by Ball’s team against the question:
“Does this give people who currently lack influence or are excluded greater ability to act?”
- https://www.lesswrong.com/posts/hcMcDfjyTMXr4EpHs/a-critique-of-dean-ball-s-the-silent-consensus
- https://www.hyperdimensional.co/p/the-silent-consensus
- https://t.me/UkhvatNews/5042
- https://openai.com/index/democratic-inputs-to-ai-grant-program-update/
- https://openaifoundation.org/news/2026-people-first-ai-fund-grantees
- https://www.nationoutside.org/our-story