As AI becomes embedded across public sector operations, how do you balance innovation with accountability? In this episode, hear expert views on the governance of AI in the public sector. Key themes include: Why defining AI clearly matters for risk management The need to keep humans in the loop for high-impact decisions Risks of using public generative AI tools with confidential data The importance of closed-loop and contextual AI systems in government How to balance centralised frameworks with agency-specific needs Building a risk-aware culture vs. enforcing tick-box compliance The challenge of maintaining transparency in machine learning Why AI governance should be principles-based, not rules-based The growing concerns around agentic AI and loss of human control Whether you're a policy maker, accountant or public service leader, this discussion offers practical insights into designing agile, ethical and effective AI governance in the public sector. Host: Ram Subramanian, External Reporting Lead, Policy, Standards and External Affairs, CPA Australia Guest: Michael Davern, FCPA Professor of Accounting and Business Information Systems, University of Melbourne. For over 30 years, both in Australia and internationally, he has led industry-engaged research projects in data analytics, business intelligence, financial reporting, risk management, data governance and ethics, among others. CPA Australia has released a report titled AI governance in the public sector. It features key insights from CPA Australia’s recent webinar with public sector experts. For more, a media release on AI governance in the public sector summarises our views. And learn more about this episode’s expert guest Michael Davern on the University of Melbourne website. Listen to more With Interest episodes and other CPA Australia podcasts on YouTube. CPA Australia publishes four podcasts, providing commentary and thought leadership across business, finance, and accounting: With Interest INTHEBLACK INTHEBLACK Out Loud Excel Tips Search for them in your podcast platform. Email the podcast team at
[email protected]