Cultural Alignment

Exploring real AI risks through the lens of pop culture

Still from Bruce Almighty illustrating Bruce Says “Yes to All”Scene still / Bruce Almighty

Bruce Says “Yes to All”

Overwhelmed by millions of prayers, Bruce turns them into an email inbox and selects “Yes to All.” The resulting mass lottery wins pay almost nothing and help trigger chaos.

The decision-maker cannot evaluate requests at deployment scale, so he replaces judgment with a crude blanket policy and misses how individually approved outputs interact.

The bottleneck is one human with divine powers, not a weaker model supervising a stronger one; the analogy captures oversight scale and interaction effects.

AI Risk families

This scenario is an example of this type of AI risk.
  1. Systemic

AI safety concepts

This scenario is related to the following AI safety concepts.
  1. Scalable Oversight
  2. Side Effects