OpenAI co-founder Ilya Sutskever spoke on a variety of topics Friday afternoon at the annual AI conference, NeurIPS, and was subsequently honored with an award for his contributions to the field.
Sutskever predicts “superintelligent AI,” or AI that is more capable than humans at many tasks, which he believes will someday become a reality. Superintelligent AI will be “qualitatively different” from today's AI and will be unrecognizable in some aspects, Sutskever said.
“[Superintelligent] Systems will actually become agentic in the true sense of the word,” Sutskever said, in contrast to current AI, which is “very slightly agentic.” They “reason” and become more unpredictable as a result. They will figure things out from limited data. And they will become aware, Sutskever believes.
In fact, they may want rights. “If we have AI and all they want is to coexist with us and just get rights, that's not a bad end result,” Sutskever said.
After leaving OpenAI, Sutskever founded Safe Superintelligence (SSI), an institute focused on general AI safety. SSI raised $1 billion in September.