Catastrophic sabotage as a major threat model for human-level AI systems

5 points | by speckx 2 days ago