AI Fiduciaries and the Law
Fiduciary duties already reach AI advisors. Proving the system honored them is the hard part, and that takes a record.
William Marks has a doctrinal piece, AI Fiduciaries and the Law, and the argument is clean: the fiduciary duties we already have reach AI advisory systems as they stand.
Care, loyalty, confidentiality. Marks walks through how each one attaches to an AI system that gives advice, and lands on a simple conclusion. You don't need new legislation to get there.
Where I'd pick it up: agreeing the duty attaches is the easy part. Proving an AI system honored it is the harder one, and that takes a record.
AI Fiduciaries and the Law
William Marks argues the existing fiduciary duties of care, loyalty, and confidentiality already reach AI advisory systems, with no new legislation required.