Speakers
‘When the Agent Gets It Wrong - Who's Accountable?’
AI agents are moving from prototype to procurement, but when something goes wrong, who is accountable? Drawing on hands-on experience building open-source governance infrastructure, this talk provides decision-makers with a practical framework for setting agent authority boundaries, creating cryptographically verifiable audit trails, and explaining why organisations that build accountability into their agent infrastructure today will survive regulatory scrutiny tomorrow.
Linda Oraegbunam is a machine learning engineer and AI governance researcher with years of experience spanning data analytics, artificial intelligence, and machine learning across the UK Civil Service and regulated commercial sectors. Her work sits at the intersection of technical delivery and institutional accountability, combining the development of production AI systems with research into the governance frameworks required to deploy them responsibly. She is the creator of a growing suite of open-source tools that form an agentic AI governance toolkit published on PyPI, and her research focuses on the relationship between agentic AI systems and organisational accountability. Linda holds an MSc in Applied Artificial Intelligence and Data Analytics from the University of Bradford and is a Fellow of the Royal Statistical Society. Linda currently serves as a reviewer for the DLI Research in Africa Days (RIAD 2026) programme and is a doctoral researcher in development focused on agentic AI governance at Leeds Beckett University. She is also the founder of AI with Linda, a public platform dedicated to making artificial intelligence practical, accessible, and relevant across industries. Through her speaking and research, Linda explores the governance responsibilities organisations must adopt as autonomous and agentic systems move from experimentation into real-world deployment.