The Human Boundary

Which forms of judgment, agency, relationship, responsibility, and authority should intelligent systems support rather than replace?

This initiative studies the boundary between assistance and delegation in social AI, medicine, education, work, firms, platforms, and government.

What changes: Technical performance is not mistaken for legitimate authority. Institutions develop explicit rules for consent, disclosure, contestability, appeal, accountability, and non-delegation.

Candidate work: Human-authority protocols and social-AI research undertaken as real studies or partner pilots, not announced as capabilities before they exist.