Sam Altman warns UN Security Council of two AI failure paths

Altman framed his remarks around a choice: AI could become "more like a new Renaissance of creativity and discovery, or more like a new Industrial Revolution of upheaval and disarray." He said he largely agrees with Professor Bengio, though he described two ways AI progress could go very badly rather than three. The first is losing control of the future to AI, where the risk is that it moves so fast people can no longer follow what is happening or intervene when needed. The second is concentration of power, where no one person, company, or country should be able to impose its worldview using the most powerful models.

He described OpenAI's approach as a middle path between blind optimism and doomerism, resting on three principles: people must remain at the center of AI decision-making; the benefits of scientific progress must be by people and for people; and the technology must empower people individually. He illustrated the pace of progress through math, from grade-school level three summers ago to a gold-medal level last summer, and said a model solved the Navier-Stokes equations a few weeks ago this summer — equations used for aircraft design, weather forecasting, and the study of blood flow.

On governance, Altman said the most important decisions cannot be made by labs in San Francisco alone and must be shaped through democratic processes and accountable governments, while companies must not substitute for the democratic process. He argued this requires international cooperation, much as countries that compete have come together before in the face of a powerful new technology, and proposed common standards so countries can compare evidence and verify compliance. Whether that cooperation materializes is likely to hinge on whether governments can agree on what good evidence, good safeguards, and good oversight look like — a question he left open by saying OpenAI does not have all the answers and is open to better ideas.