AI in Financial Services Podcast How AI Is Reshaping Regulated Professional Workflows - with Steve Hasker of Thomson Reuters
Mar 23, 2026
Steve Hasker, CEO of Thomson Reuters, a leader in legal, tax and financial information, speaks about deploying AI in regulated workflows. He discusses the need for absolute accuracy and human sign-off. He covers data isolation, transparency teams, and which regulatory filings and workflows are most likely to change first.
AI Snips
Chapters
Transcript
Episode notes
Fiduciary Work Requires Near Perfect Accuracy
- Fiduciary professionals must achieve essentially perfect accuracy when producing regulatory, transactional, or litigation work.
- Steve Hasker stresses lawyers, tax, accounting, and auditors cannot accept even small error rates because mistakes carry acute financial and legal consequences.
Define Clear Boundaries Between AI Output And Human Signoff
- Regulated industries need clear boundaries showing where machine output ends and human sign-off begins.
- Hasker says roles like general counsel or CFO must know exactly what they must sign off on versus what the AI can produce autonomously.
Customer Data Must Never Leak Into Models
- Data privacy is existential for regulated clients: customer inputs must never become part of wider AI outputs.
- Hasker contrasts Thomson Reuters' guarantee that customer inputs won't be absorbed into models with startups that often train on client data.