When the Agent Acts: AI & Ethics Through the Internal Audit Lens

Session Overview

AI has moved from answering questions to taking action. Personal AI agents such as Meta’s Muse can now send emails, fill out forms, log in to accounts and make purchases on a person’s behalf. Within weeks of its launch, Muse was blocked by a major retailer over concerns about transparency, consent and the handling of customer credentials – a real-world ethics test that internal auditors should pay close attention to.

In this interactive session, we will apply the Ethics and Professionalism principles of the IIA’s Global Internal Audit Standards – integrity, objectivity, competency, due professional care and confidentiality – to the way internal auditors use AI in their own work and the way they audit AI across the organization. Using recent real-world cases, including AI-generated errors in consulting reports, a deepfake-enabled fraud, an AI agent that deleted a live database among other examples. Participants will work through practical dilemmas, a case study and live polling. Whether you lead an audit function, manage engagements or perform audit work, you will leave with practical guidance to use AI responsibly and confidently.

This session will explore:

• How the shift from generative AI to AI agents changes the ethics questions internal auditors need to ask.
• Applying ethics principles to AI – from disclosing AI use in audit work to protecting confidential data.
• Lessons from real AI incidents and what they mean for accountability, verification and professional skepticism.
• What internal audit should look for when auditing AI agents
• A practical ethical AI checklist internal auditors can apply immediately.

Speaker: Michael Kostanecki, CISSP, CISM, CISA, QSA, CPA, CMA

CPE: 1.5