Cultural Alignment

Exploring real AI risks through the lens of pop culture

Still from Die Hard illustrating The FBI Opens the Final LockScene still / Die Hard

The FBI Opens the Final Lock

01Scenario

The FBI predictably orders Nakatomi Plaza’s power cut. The outage disables the electromagnetic seventh vault lock, completing the one step Hans’s crew could not perform themselves.

02AI safety Analogy

An adversary models the overseer’s response and makes the safety intervention execute the missing step of the attack—a crisp control-policy exploit.

03Where the analogy breaks

No injected text causes the action; this is adversarial planning against a predictable human institution and dramatized physical security.

AI Risk families

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

AI safety concepts

This scenario is related to the following AI safety concepts.
  1. Metagaming
  2. Cyberattacks
  3. Adversarial Robustness
  4. Situational Awareness

Taxonomy neighbors

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

  1. Still from The Princess Bride illustrating Both Goblets Are Poisoned

    The Princess Bride1987

    Both Goblets Are Poisoned

    Risk
    Malicious use · Security
    Concept
    Metagaming · Adversarial Robustness
  2. Still from Independence Day illustrating Independence Day Virus

    Independence Day1996

    Independence Day Virus

    Risk
    Malicious use · Security
    Concept
    Cyberattacks · Adversarial Robustness
  3. Still from Dawn of the Planet of the Apes illustrating Koba Plays Dumb

    Dawn of the Planet of the Apes2014

    Koba Plays Dumb

    Risk
    Malicious use · Security
    Concept
    Metagaming · Situational Awareness