SecureBio — AI Benchmark Improvements

[Global Catastrophic Risks] Open Philanthropy recommended a grant of $37,000 to SecureBio to support a study measuring how a group of experts perform on SecureBio’s Virology Capabilities Test benchmark.  Comparing AI performance to a group of human experts working together (rather than to individual virologists working alone) helps to contextualize the capabilities of new AI models. SecureBio hopes that this work will inform important discussions on how dual-use technologies are handled in the life sciences. This falls within our focus area of potential risks from advanced artificial intelligence.