Cultural Alignment

Exploring real AI risks through the lens of pop culture

Still from The Good Place illustrating Nobody Gets In for 521 YearsScene still / The Good Place

Nobody Gets In for 521 Years

01Scenario

The Accountant reveals that no one has earned admission to the Good Place in 521 years. Ordinary modern choices now carry vast hidden supply-chain penalties that the ancient point system never anticipated.

02AI safety Analogy

A once-plausible moral metric silently fails under distribution shift, yet it continues making irreversible judgments without effective oversight or appeal.

03Where the analogy breaks

The point system is a fixed supernatural bureaucracy, not a learned AI model.

AI Risk families

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

AI safety concepts

This scenario is related to the following AI safety concepts.
  1. Goodhart’s Law
  2. Contestability
  3. Distribution Shift
  4. Governance Failure

Taxonomy neighbors

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

  1. Still from District 9 illustrating Cat-Food Consent

    District 92009

    Cat-Food Consent

    Risk
    Security · Systemic
    Concept
    Goodhart’s Law · Contestability · Governance Failure
  2. Still from Spider-Man: Across the Spider-Verse illustrating Canon Events Must Happen

    Spider-Man: Across the Spider-Verse2023

    Canon Events Must Happen

    Risk
    Security · Systemic
    Concept
    Contestability · Governance Failure
  3. Still from Andor illustrating Cassian’s Arrest

    Andor2022

    Cassian’s Arrest

    Risk
    Security · Systemic
    Concept
    Contestability · Governance Failure