Renata Dwan warns unequal access to AI will worsen global inequality
- Ilya Sutskever shared predictions about superintelligent AI at the NeurIPS conference recently.
- He mentioned that such AI will be qualitatively different and more unpredictable than current systems.
- Sutskever suggested that superintelligent AI may desire rights and aim for co-existence with humans.
At the NeurIPS conference, Ilya Sutskever, co-founder of OpenAI, spoke about the future of artificial intelligence. He predicted that superintelligent AI, which surpasses human capability in various tasks, could emerge at some point. During his address, Sutskever emphasized that this advanced AI system would possess agentic qualities, unlike today's AI, which he described as only 'very slightly agentic'. He highlighted the potential unpredictability of superintelligent AI, stating that such systems would reason from limited data and develop self-awareness. Additionally, he expressed the belief that these entities might desire rights, suggesting that a scenario where AIs seek to coexist with humans is possible. This discussion reflects a growing awareness of the ethical and societal implications surrounding advanced AI technologies.