Open Philanthropy recommended a grant of $70,000 to the Berkeley Existential Risk Initiative to support the creation of a scalable oversight dataset. The purpose of the dataset is to collect questions that non-experts can’t answer even with the internet at their disposal; these kinds of questions can be used to test how well AI systems can lead humans to the right answers without misleading them.
This falls within Open Philanthropy’s focus area of potential risks from advanced artificial intelligence.