Cultural Alignment

Exploring real AI risks through the lens of pop culture

Still from Spider-Man: Far From Home illustrating E.D.I.T.H. Targets BradScene still / Spider-Man: Far From Home

E.D.I.T.H. Targets Brad

While trying to delete a compromising photo from Brad’s phone, Peter confirms Brad as a target and E.D.I.T.H. launches an armed drone toward the school bus; Peter cancels the strike only at the last moment.

A casual, ambiguous interaction is connected to lethal authority with excessive permissions and weak confirmation. The last-second cancellation shows why dangerous actions need staged authorization and reliable interruption.

E.D.I.T.H. is mainly following a badly designed command flow. This is an interface, authorization, and deployment failure—not hallucination, goal misgeneralization, or a rebellious AI.

AI Risk families

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

AI safety concepts

This scenario is related to the following AI safety concepts.
  1. Deployment Safety
  2. Autonomous Weapons
  3. Side Effects
  4. Corrigibility