Cultural Alignment

Exploring real AI risks through the lens of pop culture

Still from Spider-Man: No Way Home illustrating Changing the Spell Mid-CastScene still / Spider-Man: No Way Home

Changing the Spell Mid-Cast

Doctor Strange begins a world-scale memory spell before Peter has settled the requirements. Peter adds exception after exception mid-cast; Strange warns him to stop changing the parameters, then the spell breaks containment and opens the multiverse.

A high-impact operation starts before requirements are elicited, then accepts live specification changes without a clean abort-and-restart path. It is a vivid case for change control, rollback, and staged deployment.

The mechanism is magic, and the main failure is a reckless human workflow rather than autonomous optimization.

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. Side Effects
  3. Outer Alignment
  4. Corrigibility