Social Contract for Humans and Agents

Framing

This piece turns from productivity to legitimacy. If agents become durable collaborators in messaging systems, workplaces, and consumer products, the old default trust assumptions are no longer enough. The blog post will ask what rights, obligations, and oversight mechanisms should govern human-agent relationships.

Initial Claims

• Delegation to agents must come with explicit boundaries and a clear right of revocation.
• Trust in agents depends on controllability, recoverability, and auditability, not only capability.
• The future social contract will involve humans, agents, and the platforms that mediate them.

Draft Outline

1. Why capability alone is not enough
2. Delegation, authority, and revocation
3. What platforms owe to users
4. Why trustworthiness matters more than benchmark theater
5. Toward a practical social contract for agentic systems