Summary
Researchers from major universities are concerned about certain biological data being available online because it could help AI create dangerous viruses. They propose treating this sensitive data with the same care as private health information to ensure that it is not misused. This move comes amidst the U.S. government’s efforts to speed up AI development for scientific research.
Key Facts
- Researchers from Johns Hopkins, Oxford, Stanford, Columbia, and NYU are involved in this effort.
- Over 100 researchers have endorsed a framework to regulate specific biological data.
- The concern is that AI trained on certain viral genetics data could design harmful viruses.
- The White House's Genesis Mission aims to use AI and big datasets to boost research.
- Some developers choose not to include virology data in AI training due to misuse risks.
- There currently isn't clear guidance on which biological data should be restricted.
- Researchers emphasize responsible data sharing, allowing access while protecting from misuse.
- The framework suggests governments should continually review data restrictions as science advances.
