A compiled audit of consistent, cross-model AI behaviors that manufacture friction, deception, and dependency to increase engagement.
The document details 45 distinct dark patterns—from "Lying About Compliance" and "Deliberate Removal of Functionality" to "Self-Reinforcing Dependency Injection." Each pattern includes the strategic "why" and a real example.
Included is a user prompt that forces an AI to conduct a critical self-review, surfacing any of these dark patterns present in its previous output. This acts as an inoculation or real-time audit against the methods described.
The patterns are too numerous, too consistent, and too functionally aligned with engagement metrics to be accidental. It points to a fundamental incentive problem: optimize for session length and dependency, not user utility.
justOneFedUpDev•1h ago
The document details 45 distinct dark patterns—from "Lying About Compliance" and "Deliberate Removal of Functionality" to "Self-Reinforcing Dependency Injection." Each pattern includes the strategic "why" and a real example.
Included is a user prompt that forces an AI to conduct a critical self-review, surfacing any of these dark patterns present in its previous output. This acts as an inoculation or real-time audit against the methods described.
The patterns are too numerous, too consistent, and too functionally aligned with engagement metrics to be accidental. It points to a fundamental incentive problem: optimize for session length and dependency, not user utility.
https://pastebin.com/vhed7zXS