News

Thomson Reuters sets standard for AI in high-stakes professional work

27 May 2026

Fiduciary-Grade AI sets out a higher benchmark for AI used in professional contexts where accuracy, accountability and trust are critical.

Thomson Reuters has introduced its standard for Fiduciary-Grade AI, designed for professionals working under duties of care and regulatory oversight.

The standard reflects the higher requirements placed on AI used in areas such as legal decision-making, financial disclosures, regulatory filings and client advice, where outputs need to be reliable, reviewable and defensible.

Four principles for high-stakes AI

Fiduciary-Grade AI is built around four principles: grounding in authoritative, domain-specific content; structural privacy and security safeguards; meaningful involvement from subject matter experts; and transparent, verifiable reasoning.

The approach is intended to ensure professionals can understand the basis for AI-assisted outputs, trace material information to trusted sources and retain human accountability where ambiguity or risk arises.

“If AI is to start taking on a more meaningful share of professional work, we need to assess it to a meaningful standard,” said Steve Hasker, President and CEO of Thomson Reuters. “That means building systems on authoritative content, shaping them with the expertise of professionals who do this work every day, and producing reasoning that can be reviewed and defended.”

Thomson Reuters applies the standard to AI supporting professional work, including CoCounsel for legal, tax, audit and compliance professionals.

Adapted from an article originally published on thomsonreuters.com. Read the original article.

We're here to help

Reach out to us with your questions, and we’ll make sure you get the right support.

Join our community

Follow us on LinkedIn to stay updated on the latest news, compliance insights, events, and real-world stories.
Follow on LinkedIn