Thomson Reuters Launches Fiduciary-Grade AI Standard for Professional Workflows
TRI•Thomson Reuters introduced its Fiduciary-Grade AI standard, a benchmark requiring outputs to be grounded in authoritative, domain-specific content with traceable sources and built-in privacy and security safeguards. It underpins CoCounsel for legal, tax, audit and compliance professionals with expert-shaped workflows, real-time human support and transparent reasoning for high-stakes work.
1. Introduction of Fiduciary-Grade AI Standard
Thomson Reuters introduced its Fiduciary-Grade AI standard as a higher benchmark for AI in professional contexts requiring accuracy, accountability and traceable outputs. The standard mandates that every AI output derive from authoritative, curated content and that systems include structural privacy and security safeguards.
2. Four Defining Principles
Fiduciary-Grade AI is defined by four principles: grounding in domain-specific content with traceable sources; built-in data privacy and security; workflows designed and refined with credentialed experts; and transparent, verifiable reasoning that surfaces each step for professional review.
3. Integration with CoCounsel and Client Impact
This standard underpins the CoCounsel suite for legal, tax, audit and compliance professionals, requiring real-time human support and expert-shaped workflows. By ensuring outputs can be independently verified and defended, it aims to drive deeper AI integration in regulated, high-stakes environments.
