Over 100 scientists have signed an agreement to prevent the misuse of their AI-supported research in the development of biological weapons. The agreement comes after an AI experiment generated 40,000 hypothetical bioweapons in just six hours. The AI system used, MegaSyn, was inverted from its usual operation of detecting toxicity in molecules to avoid them, to instead retain toxic molecules. The experiment serves as a warning of the potential dangers of AI misuse. The researchers have kept some details of their research secret due to the sensitive nature of the findings.
Read more at: www.nytimes.com