OpenAI CEO Altman: AI Moves From Tool to Discoverer, But Must Stay Human-Centered

OpenAI chief executive Sam Altman said during an in-depth interview in October 2026 that AI is shifting from a tool to a "discoverer," helping humans explore entirely new knowledge — but the future must remain human-centered. He also warned that if recursive self-improvement (RSI) is strongly pursued, it could happen, but continuing down that path would be very dangerous.

Altman said AI has already demonstrated its ability to carry out complex research in mathematics, and he expects the same approach to work in biology, drug discovery, materials science and new energy. He hopes to see suitable drug candidates for major diseases within a few years and said he is personally excited about the prospect of a major physics discovery.

On personal AI assistants, Altman said one of the most exciting developments of 2026 is that the personal AI assistant revolution began with OpenClaw. AI can perform tasks on behalf of users in various environments, but more product work and even new consumer hardware are needed. He noted that laptops and phones were designed for humans rather than for AI, and that AI needs new form factors to support ambient awareness and more powerful computing.

Regarding machine consciousness, Altman said today's models are not conscious and should not be assigned personality. He expressed concern that debates over AI consciousness could be used to argue that humans should not always be in control, and stressed that AI is a resource for humanity and the future must be human-centered.

On compute demand, Altman said that if half the world's population fully used personal AI assistants, the need for computing power could not be measured with a static market size; "as many GPUs as we can get" was his answer. He also said OpenAI's core advantage lies in being able to deploy next-generation models earlier than others and use that internal lead to build new things.