AI agency requires more than safety guardrails: online interactions are lopsided negotiations that need collective bargaining mechanisms built into agentic systems.
Key Takeaways
The author frames every online interaction as a negotiation where users currently have little structural leverage against platforms and services.
“Agentic AI” as typically sold focuses on task execution and safety, but skips the question of who the agent is actually bargaining for.
True user agency means AI systems that represent user interests in adversarial or misaligned environments, not just assistants that complete tasks.
Collective bargaining as a metaphor suggests the solution is systemic and architectural, not just a matter of individual model alignment or prompt design.