OpenAI's Plan to Control Superhuman AI: Demos and Alignment

OpenAI's Superalignment team is working on developing ways to control and govern "superintelligent" AI systems that surpass human intelligence. Led by OpenAI co-founder and chief scientist Ilya Sutskever, the team aims to build governance and control frameworks for future powerful AI systems. They are using a weaker AI model to guide a more advanced model in desirable directions. OpenAI is also launching a $10 million grant program to support research on superintelligent alignment and plans to host an academic conference on superalignment in 2025. The involvement of former Google CEO Eric Schmidt, who is funding a portion of the grant, raises questions about the availability and use of OpenAI's research.