ABSTRACT
This chapter examines how artificial intelligence reshapes fiduciary law by technologically mediating the exercise of entrusted discretion. Rather than displacing the traditional foundations of fiduciary obligation – entrustment, vulnerability, discretion, loyalty, and care – AI redistributes the performance of fiduciary functions across complex socio-technical chains involving fiduciaries, institutions, models, developers, providers, data, and beneficiaries. The analysis distinguishes four configurations: AI as an instrument of fiduciary decision-making, as a de facto delegate of fiduciary judgement, as a putative fiduciary, and as the object of fiduciary-type obligations imposed on those who design, provide, and control AI systems. Across these configurations, AI alters the content of duties of care, supervision, and informed judgement; complicates attribution where human decision-making becomes largely ratificatory; and generates architectural conflicts of interest that conventional disclosure-and-consent mechanisms may struggle to address. The chapter further highlights the recursive role of AI as both an object and an instrument of fiduciary monitoring. It argues that these developments converge on a principle of non-displacement of fiduciary accountability: although fiduciary functions may migrate to machines and technological infrastructures, responsibility for the loyal and careful exercise of entrusted power must remain attributable to legally cognisable actors.
Passador, Maria Lucia, AI and Fiduciary Law (August 16, 2026) in D Gordon Smith and Y Naftali Ben Zion (eds), Concise Encyclopedia of Fiduciary Law (Edward Elgar Publishing, forthcoming).
Leave a Reply