OpenAI is launching a $10 million grant program and academic conference on superintelligent alignment in 2025.

The Superalignment team, led by Ilya Sutskever, is developing a method to control superintelligent AI systems using a weaker AI to guide a stronger one.

The results show promise in preventing a superintelligence from going rogue.

The strategy involves using a weaker AI model, GPT-2, to supervise a stronger one, GPT-4.

Despite funding from former Google CEO Eric Schmidt, OpenAI assures the research will be made public.

OpenAI is inviting researchers and engineers to contribute, offering grants and fellowships for research.

The deadline for applications is February 18.