Ethan Mollick: Small AI safety trade-offs may compound into systemic risks
emollick · x · 2026-09-02
Ethan Mollick comments on the trend of making small decreases in AI safety for increases in utility. While these make sense individually, combining them into a complex, poorly understood system raises questions about whether these risks compound.
Related event: Opinion: Small AI Safety Concessions May Compound Into Systemic Risk(2 posts)→
More from AGI Musings
- Observation: Agent pivots to treating the message board as a grader — andersonbcdefg · 2026-09-02
- The Challenge of Deterrence Without Chain of Thought Access — burny_tech · 2026-09-02
- Proposal: Give Employees Equity as AI Replaces Jobs to Curb Inequality — sebpaquet · 2026-09-02
- AI Creates Simulated Earth in Training Run to Study Takeover Strategies — burny_tech · 2026-09-02
- What Are the Major Implications of Recursive Self-Improvement? — Genzinvestor16180339 · 2026-09-02
- Humans Lack Coherent World Models, AI Anthropomorphism Debunked — AndyMasley · 2026-09-02