AI Experts Call for Evidence-Based Debate on Artificial Intelligence Risks
Summary
AI experts at the United Nations have urged an evidence-based discussion of artificial intelligence risks, warning against exaggerated fear-driven narratives. Speakers including Joelle Barral, Maria Ressa and Yoshua Bengio said the debate should focus on what is known, what is uncertain and how far risks can reasonably be extrapolated.
Key Points
- Experts linked to the UN’s international scientific panel on AI called for scientific evidence and reasoning to guide debate on AI risks.
- Joelle Barral said fear-based narratives are not useful and stressed the importance of distinguishing knowns from unknowns in AI research.
- Maria Ressa argued the AI debate should not be reduced to extremes and said the reality is more nuanced and complex.
- Yoshua Bengio said scientists must be careful about what they can know from evidence and what they can only extrapolate.
