Recent advancements in AI technology could significantly heighten the risk of human-caused pandemics, with experts predicting a fivefold increase compared to last year. A study by the Forecasting Research Institute revealed that if AI systems, capable of outperforming PhD-level virologists, assist novice biologists in creating bioweapons, the estimated annual risk would rise to 1.5%. Major AI firms like OpenAI and Anthropic have expressed concerns about AI aiding malicious actors. Potential mitigations include implementing safeguards at the model level to prevent AI from producing harmful content and enforcing regulations on genetic material synthesis to ensure safety. Current laws allow companies to synthesize genetic codes without screening, raising biosecurity risks. Experts suggest that with proper safeguards, the risk could return to a manageable 0.4%. This emerging threat underscores the need for robust policies to tackle AI-related biohazard risks.
Source link
Share
Read more