Cultural Alignment
Still from Ready Player One illustrating 80% Before Seizures

80% Before Seizures

Sorrento proposes filling up to 80% of an OASIS user’s visual field with advertising—the estimated limit before seizures.

Human harm is treated as a boundary to saturate while revenue remains the true objective. An optimizer following that policy will push every user to the edge of the safety constraint rather than preserve their welfare.

The proposal is made by a human executive, so the analogy is to incentive design and optimization targets rather than autonomous AI intent.

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. Goodhart’s Law
  2. Side Effects
  3. Governance Failure