---
type: "synthesis"
theme: "fiduciary-duty"
sources: ["governance"]
id: "cross-fiduciary-thread"
sourceVaultSlug: "hbr-seg-governance"
originDay: 7
articleStem: "hbr-seg-governance"
sourceUrl: "(unified vault: 8 sources)"
sourceTitle: "HBR — Firm Ⅱ-B · Governance, decision rights, leadership, risk"
---
# Fiduciary Duty as Connective Tissue

The oldest legal concept in the corpus quietly binds its newest problems. Fiduciary duty — the highest standard of loyalty and care ([[prereq-fiduciary-duty]]) — appears in four guises.

- **The board's fiduciary duty** is the baseline assumption of the two board articles: [[prereq-board-fiduciary-duties]] and [[prereq-corporate-governance-d7]]. *Decision-Making by Consensus* weaponizes it — [[claim-boards-failing-governance]] argues that accepting filtered reports is *failing* that duty.
- **The agent's fiduciary duty** is *Can AI Agents Be Trusted?*'s central proposal: [[concept-ai-fiduciary-duty]] and [[action-establish-ai-fiduciary-status]] would bind autonomous software to loyalty, disclosure, and care, with [[claim-fiduciary-legal-precedent]] suggesting existing precedent may already reach it.
- **The unassignable duty** is the shared limit. *How C-Suite Roles Are Reshaped* runs [[concept-agentic-governance]] straight into it — [[question-ai-accountability-d7]] — because corporate law presumes *human* directors with fiduciary duties, so an AI "board member" faces a legal ceiling, not just a technical one. [[question-enforcing-ai-fiduciary-duty]] echoes: fiduciary duties attach to persons and institutions, not code.

The synthesis: fiduciary duty is the corpus's mechanism for re-attaching the [[cross-single-owner-principle|single owner]] to a responsible legal person whenever automation threatens to dissolve accountability. Its enrichment caveat is uniform: the accountable party is realistically the developer/deployer/organization, never the software itself.