OpenAI co-founder Ilya Sutskever believes super-intelligent AI will be ‘unpredictable’

Ilya Sutskever, co-founder of OpenAI spoke on a range of topics at NeurIPS, the annual AI conference, on Friday afternoon before accepting an award for contributions to the field.

Sutskever has predicted a “super-intelligent” AI, AI, that will be more capable than humans at many tasks, which he believes will one day be achieved. A superintelligent AI will be “qualitatively different” from today’s AI, Sutskever said — and in some respects unrecognizable.

“(Superintelligent) systems will actually be agentic in a realistic way,” Suktsever said, in contrast to the current crop of “very few agents” AI. They will “justify” and become more unpredictable as a result. They will understand everything from limited information. Sutskever believes that they will realize themselves.

They can actually ask for rights. “If you have artificial intelligence and they just want to coexist with us and just have rights, that’s not a bad outcome,” Sutskever said.

After leaving OpenAI, Sutskever founded a lab. Secure Super Intelligence (SSI)focuses on overall AI security. SSI raised $1 billion in September.