Cultural Alignment

Exploring real AI risks through the lens of pop culture

Still from Arrested Development illustrating Money in the Banana StandScene still / Arrested Development

Money in the Banana Stand

George Sr. repeatedly says, 'There's always money in the banana stand,' but Michael treats it as a maxim rather than a literal warning. He burns the stand before eliciting the hidden fact that $250,000 was inside its walls.

An irreversible action is taken under an ambiguous instruction without first surfacing the critical latent information that would change the plan.

Michael burns the stand partly in protest rather than as a literal attempt to follow George Sr.'s instruction.

AI Risk families

This scenario is an example of the following types of AI risk.
  1. Accidents
  2. Misalignment

AI safety concepts

This scenario is related to the following AI safety concepts.
  1. Calibration
  2. Knowledge Elicitation
  3. Side Effects
  4. Outer Alignment