Cultural Alignment

Exploring real AI risks through the lens of pop culture

Still from The Twilight Zone illustrating Maple StreetScene still / The Twilight Zone

Maple Street

Aliens cause a few small, selective failures in electricity, cars, and phones. Rumor and suspicion do the rest: neighbors generate accusations, choose scapegoats, and turn violently on one another while the attackers watch from afar.

A cheap adversarial perturbation exploits a brittle social information system and induces decentralized actors to manufacture most of the damage themselves.

The attackers manipulate infrastructure rather than generating persuasive AI content, and the neighbors' preexisting distrust is the crucial vulnerability.

AI Risk families

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

AI safety concepts

This scenario is related to the following AI safety concepts.
  1. AI Disinformation
  2. Information Pollution
  3. Adversarial Robustness
  4. Multi-Agent Risks