Cultural Alignment

Exploring real AI risks through the lens of pop culture

Still from Fargo illustrating Machines Can’t See GloriaScene still / Fargo
Archive noteNo clip in the collectionThe scene analysis remains available in full.

Machines Can’t See Gloria

  • Fargo
  • The Principle of Restricted Choice

01Scenario

Automatic doors, faucets, hand dryers, and her phone repeatedly fail to register Gloria, making her wonder aloud whether she exists.

02AI safety Analogy

People outside a system’s sensing assumptions can become functionally invisible even when the system appears to work for everyone else.

03Where the analogy breaks

The failures are a surreal character motif rather than an explicit biased algorithm.

AI Risk families

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

AI safety concepts

This scenario is related to the following AI safety concepts.
  1. Algorithmic Bias
  2. Contestability
  3. Distribution Shift

Source continuation

More from Fargo

View all 2 scenarios
  1. Still from Fargo illustrating Varga’s Fake Serial Killer

    Fargo2017

    Varga’s Fake Serial Killer

Taxonomy neighbors

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

  1. Still from The Pianist illustrating The German Coat Classifier

    The Pianist2002

    The German Coat Classifier

    Risk
    Accidents · Systemic
    Concept
    Contestability · Distribution Shift
  2. Still from Chappelle's Show illustrating Clayton Bigsby’s World Model

    Chappelle's Show2003

    Clayton Bigsby’s World Model

    Risk
    Systemic
    Concept
    Algorithmic Bias · Distribution Shift
  3. Still from The Good Place illustrating Nobody Gets In for 521 Years

    The Good Place2019

    Nobody Gets In for 521 Years

    Risk
    Systemic
    Concept
    Contestability · Distribution Shift