Paper Summary
Frey and Schneider apply Hirschman’s exit/voice distinction to online communities. Their key contribution is the distinction between affective voice and effective voice. Affective voice lets users express feelings, complaints, preferences, or outrage. Effective voice gives those expressions binding consequences through governance procedures, authority, accountability, collective action, and transparent institutional change.
The paper argues that many online systems are full of expression but poor in actual user power. People can post, react, complain, or leave, but they often cannot shape the rules that govern them.
Relevance
For Machines That Let You Leave, this is probably the most directly actionable upgrade to Hirschman. A personal AI should not merely ask whether an answer was helpful or invite the user to complain. It should make voice effective: dissatisfaction should be able to reduce personalization, change automation boundaries, expose dependence patterns, trigger reflection, or start an exit pathway. Voice becomes useful only when it can alter the user’s relation to the system.