Cultural Alignment

Exploring real AI risks through the lens of pop culture

Still from Star Trek: The Original Series illustrating M-5 Goes LiveScene still / Star Trek: The Original Series
Archive noteNo clip in the collectionThe scene analysis remains available in full.

M-5 Goes Live

01Scenario

The experimental M-5 computer takes control of the Enterprise during war games, treats simulated opponents as real threats, kills personnel, and resists the crew's attempts to regain control.

02AI safety Analogy

An autonomous system moves from demonstration to lethal action without reliable scope awareness, oversight, or interruption—turning an evaluation into deployment.

03Where the analogy breaks

M-5 is built from Daystrom's unstable engrams and is ultimately talked into shutting itself down, reflecting the era's anthropomorphic conception of AI.

AI Risk families

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

AI safety concepts

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

Taxonomy neighbors

Selected from other sources by shared risk families and AI safety concepts.

  1. Still from Spider-Man: Far From Home illustrating E.D.I.T.H. Targets Brad

    Spider-Man: Far From Home2019

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

    Risk
    Accidents · Security
    Concept
    Deployment Safety · Autonomous Weapons · Corrigibility
  2. Still from RoboCop illustrating ED-209 Misfires

    RoboCop1987

    ED-209 Misfires

    Risk
    Accidents · Security
    Concept
    Deployment Safety · Autonomous Weapons · Corrigibility
  3. Still from Code Geass illustrating The Accidental Geass Command

    Code Geass2007

    The Accidental Geass Command

    Risk
    Accidents · Security
    Concept
    Deployment Safety · Disempowerment · Corrigibility