OpenAI author rejects claim that AI will inevitably end human primacy

  • OpenAI
  • silent consensus
  • human primacy

Article explores future of artificial intelligence following Black October

Author challenges inevitability of machine control over global affairs

An OpenAI employee publicly rejected what the essay calls the “silent consensus”: the belief that increasingly capable AI will inevitably displace humans as the leading force in world affairs. OpenAI published the essay, which its author says is adapted from a lecture in Harvard’s Vik-Bailey Lecture series; the author’s views do not necessarily represent OpenAI or its other employees.

The author argues that people should retain authority over institutions and economic activity as AI systems take on more work. The essay warns that separate decisions to delegate tasks to machines could collectively leave people with less influence, even without an AI rebellion or an intention by its owners to disempower humans.

To preserve human accountability, the author calls for technical safeguards that connect people to the actions of their AI agents. The essay proposes limiting agents’ permissions, making their actions auditable and allowing users to withdraw their authority. It also urges extraordinary caution about allowing agents to form corporations, own property or litigate, if such powers are permitted at all.

The author also advocates building AI systems compatible with human primacy, including tools that help people understand issues and make decisions. The essay presents these ideas as starting points, not a complete plan for a society with superhuman AI.