Anirudh Veeraragavan on Defensible Financial Agents

AI Engineer13m 28s
Anirudh Veeraragavan in a light-blue T-shirt beside the headline Defensible Financial Agents on black.
0 comments · 0 votesOpen discussionClose discussion
Sign in to join the discussion

    Anirudh Veeraragavan uses a bank reconciliation audit to explain why strong aggregate evaluation scoresEvaluation is the systematic assessment of a system's behaviour or results against explicit criteria. do not establish how a particular agent decision was made.

    Anirudh Veeraragavan describes state-aware authorizationAn agent permission boundary defines and enforces which resources and actions an automated agent is allowed to use., execution records and evidence tied to a specific document and version. These controls help prevent changes to reviewed work and allow individual steps to be examined.

    Anirudh Veeraragavan discusses errors in evaluation data and situations that should trigger human reviewA risk assessment identifies ways something could cause harm and examines their likelihood and consequences under specified conditions.. He applies the same decision-assurance principles to other financial workflows.

    Original YouTube thumbnailWatch on YouTube

    Video topic coverage

    • Agents & workflows100
    • Coding & tools36.2
    • Models & learning0
    • Infrastructure & compute0
    • Research & evaluation63.8
    • Safety & security95.7
    • Business & productivity91.5
    • Creative & media0
    • Robotics & perception0
    • Society & policy0

    Share this video

    Anirudh Veeraragavan in a light-blue T-shirt beside the headline Defensible Financial Agents on black. With WWW.ARTIFICIAL-INTELLIGENCE.VIDEO, 7 November 2026 and duration 13m 28s.

    Anirudh Veeraragavan explains controls that help people review and defend AI-assisted financial decisions.