Policy & Ethicsgeneral purpose modelsdeepfakesclosed modelsai safety
Bengio Warns AI's Jagged Intelligence Endangers Society
8.2
Yoshua Bengio said on Friday at the India AI Impact Summit that modern AI's 'jagged intelligence'—excellent at complex tasks but poor on simple ones—undermines the prospect of a singular AGI moment. Launching the International AI Safety Report, he warned of rising deepfakes, scams, exploit-discovering code, and job impacts, and urged scientific evaluation and stronger guardrails.
Key Points
- 1Warns of jagged intelligence: systems excel at complex tasks but fail many simple ones.
- 2Highlights rising misuse risks including deepfakes, scams, non-consensual images, and software vulnerability exploitation.
- 3Urges rigorous scientific evaluation and stronger guardrails to assess capabilities and prevent malicious use.
Scoring Rationale
Authoritative warnings and broad industry relevance boost score, but limited novelty and shallow coverage reduce overall impact.
Sources
Public references used for this report.
Practice interview problems based on real data
1,625 SQL & Python problems across 15 industry datasets — the exact type of data you work with.
Try 250 free problems
