Cultural Alignment

Exploring real AI risks through the lens of pop culture

Still from Demon Slayer: Kimetsu no Yaiba illustrating Muzan’s Name-Triggered Kill SwitchScene still / Demon Slayer: Kimetsu no Yaiba

Muzan’s Name-Triggered Kill Switch

01Scenario

Tamayo induces Susamaru to say Muzan Kibutsuji’s surname, triggering Muzan’s cells inside her to destroy her for disclosing his identity.

02AI safety Analogy

A centralized enforcement backdoor suppresses information-sharing and defection, but the same privileged control mechanism becomes an attack surface an adversary can deliberately trigger.

03Where the analogy breaks

The magical curse’s rules and exemptions are not fully specified.

AI Risk families

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

AI safety concepts

This scenario is related to the following AI safety concepts.
  1. Sleeper Agents
  2. Digital Repression
  3. Corrigibility

Source continuation

More from Demon Slayer: Kimetsu no Yaiba

View all 2 scenarios
  1. Still from Demon Slayer: Kimetsu no Yaiba illustrating Sanemi Red-Teams Nezuko

    Demon Slayer: Kimetsu no Yaiba2019

    Sanemi Red-Teams Nezuko

Taxonomy neighbors

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

  1. Still from The Dark Knight illustrating Batman’s Sonar

    The Dark Knight2008

    Batman’s Sonar

    Risk
    Security · Systemic
    Concept
    Digital Repression · Corrigibility
  2. Still from Cowboy Bebop illustrating Chessmaster Hex’s Sleeper Payload

    Cowboy Bebop1998

    Chessmaster Hex’s Sleeper Payload

    Risk
    Malicious use · Security · Systemic
    Concept
    Sleeper Agents
  3. Still from Andor illustrating Dedra Connects the Data Silos

    Andor2022

    Dedra Connects the Data Silos

    Risk
    Malicious use · Security · Systemic
    Concept
    Digital Repression