OECD Dark Pattern Harm Taxonomy

The OECD framework for categorizing deceptive design patterns and their harms. Provides systematic vocabulary for discussing manipulation in digital interfaces.

Pattern Categories

Forced Action: Requiring unrelated actions (forced account creation) Interface Interference: Manipulating UI to guide choices (hidden options) Nagging: Persistent requests to change behavior Obstruction: Making desired actions difficult (hard to unsubscribe) Sneaking: Hidden information or actions (added fees at checkout) Social Proof: Fake or misleading popularity signals Urgency: False scarcity or time pressure

Harm Categories

Financial: Direct monetary costs Privacy: Data exposure beyond expectations Autonomy: Undermining informed choice Time: Wasted attention and effort Psychological: Stress, confusion, manipulation

AI Relevance

AI systems can perpetuate dark patterns through:

• Recommendation systems that maximize engagement over user benefit
• Persuasion techniques embedded in conversational AI
• Personalized manipulation based on user data
• Opacity that prevents informed consent

Related: 07-molecule—ui-as-ultimate-guardrail, 01-atom—configuration-as-design