OpenAI co-founder Ilya Sutskever spoke On Friday afternoon, he spoke on the annual NeurIPS AI conference on a variety of topics before accepting an award for his contributions to the sphere.
Sutskever made his predictions for a “superintelligent AI” – one which is more capable than humans at many tasks and which he believes will eventually be achieved. Superintelligent AI can be “qualitatively different” than the AI we’ve got today, Sutskever said – and unrecognizable in some facets.
“(Superintelligent) systems will actually be agentic in a possible way,” Sutskever said, unlike the present generation of “very barely agentic” AI. They change into “superior” and due to this fact more unpredictable. You will understand things based on limited data. And they can be self-aware, Sutskever believes.
Maybe they really want rights. “It's not a foul final result if there are AIs they usually just wish to coexist with us and just have rights,” Sutskever said.
After leaving OpenAI, Sutskever founded Safe Superintelligence (SSI), a lab focused on general AI safety. SSI raised $1 billion in September.
