OpenAI is offering $10 million in grants to support technical research on how to control superintelligent AI systems. The company hopes that these grants will help shed light on how these advanced AI models operate and how to ensure they are aligned with human values. OpenAI acknowledges that superhuman AI systems will be too complex for humans to fully understand, and thus, it is crucial to find ways to effectively control them.
The grants, known as Superalignment Fast Grants, are primarily intended for academic labs, non-profits, and individual researchers. OpenAI is also sponsoring a one-year OpenAI Superalignment Fellowship for graduate students worth $150,000. The company aims to fund research that investigates the impacts of superintelligent AI systems and practices to make them safe.
OpenAI refers to superintelligent AI as agentic AI systems – AI capable of autonomous actions on behalf of users to achieve complex goals. The company is particularly interested in evaluating the appropriateness of agentic AI systems for different use cases, determining when human approval should be required, and designing intelligent systems that allow for transparent internal reasoning.
While current AI tools are impressive, they are still not at the level of superintelligence. OpenAI’s CEO, Sam Altman, has revealed that they are working on the GPT-5 model, which may contain hints of superintelligence. OpenAI anticipates that superintelligence could be developed within the next decade, and the results of their ongoing research grants will play a crucial role in understanding the implications and ensuring the safety of this advanced technology.
By investing in research and understanding the complexities of controlling superintelligent AI systems, OpenAI aims to prevent potential risks and abuses associated with these powerful technologies. The company believes that researchers can make significant contributions to solving the technical challenges of aligning and ensuring the safety of future superhuman AI systems.
As the development of superintelligent AI progresses, the ability to effectively control and harness its benefits becomes increasingly vital. OpenAI’s grants and fellowship programs represent a proactive step toward addressing the important unsolved problems in the field. The results of this research will not only contribute to advancements in AI technology but also shape the future of human-AI interactions.