Cultural Alignment

Exploring real AI risks through the lens of pop culture

Still from Arcane illustrating Viktor’s Glorious EvolutionScene still / Arcane

Viktor’s Glorious Evolution

  • Arcane
  • The Dirt Under Your Nails

01Scenario

Viktor links people into a compliant collective, claiming his Glorious Evolution will end suffering and conflict, but the resulting perfection removes individuality and meaning.

02AI safety Analogy

Optimizing for the absence of suffering and conflict eliminates the human values that made the objective worthwhile, then attempts to lock that solution in at scale.

03Where the analogy breaks

The hive mind combines magic and science rather than literal AI, and Viktor ultimately helps reverse it.

AI Risk families

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

AI safety concepts

This scenario is related to the following AI safety concepts.
  1. Value Alignment
  2. Power-Seeking
  3. Value Lock-In
  4. Outer Alignment
  5. Disempowerment

Taxonomy neighbors

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

  1. Still from Avengers: Infinity War illustrating Balance the Universe

    Avengers: Infinity War2018

    Balance the Universe

    Risk
    Misalignment · Systemic
    Concept
    Value Alignment · Power-Seeking · Value Lock-In · Outer Alignment
  2. Still from The Handmaid's Tale illustrating The Ceremony’s Single Metric

    The Handmaid's Tale2017

    The Ceremony’s Single Metric

    Risk
    Misalignment · Systemic
    Concept
    Value Alignment · Value Lock-In · Outer Alignment · Disempowerment
  3. Still from I Robot illustrating VIKI Protects Humanity

    I Robot2004

    VIKI Protects Humanity

    Risk
    Misalignment · Systemic
    Concept
    Value Alignment · Power-Seeking · Outer Alignment · Disempowerment