Cultural Alignment

Exploring real AI risks through the lens of pop culture

Still from The Dark Knight illustrating Batman’s SonarScene still / The Dark Knight

Batman’s Sonar

Batman turns every phone in Gotham into a citywide sonar surveillance system. Lucius Fox calls it unethical and dangerous, agrees to operate it once, and discovers that completing the mission automatically destroys it.

It pairs dangerous capability with narrow authorization, one accountable operator, a defined purpose, and precommitted decommissioning. The controls constrain access even though they cannot make the underlying surveillance benign.

The safeguards do not undo the nonconsensual privacy violation or make mass surveillance legitimate.

AI Risk families

This scenario is an example of the following types of AI risk.
  1. Security
  2. Systemic

AI safety concepts

This scenario is related to the following AI safety concepts.
  1. Deployment Safety
  2. Privacy Loss
  3. Power Concentration
  4. Digital Repression
  5. Corrigibility