Dr. Karthik Vaidhyanathan, Assistant Professor at IIIT Hyderabad, participated in a panel discussion at FinAI Confluence 2026 titled “Trust the Agent? Governing AI in Regulated Financial Services.” The panel explored an increasingly important question as AI agents become more autonomous: Can we trust them to operate within the boundaries we set?

Dr. Karthik highlighted several key considerations for building trustworthy Agentic AI systems. These included designing autonomy according to the risk associated with a task, managing uncertainty, strengthening observability beyond simply evaluating an agent’s output, and continuously assessing agent behaviour before and after deployment.

He emphasised that governance mechanisms such as access controls, monitoring, guardrails, decision limits, and mechanisms to safely stop agents need to be considered as part of the overall system architecture. Trust, he noted, needs to be established progressively before autonomy is scaled.

The discussion brought together experts from academia and industry to examine the challenges of deploying Agentic AI in regulated financial services, with a focus on governance, accountability, responsible AI, and trust.

Dr. Karthik’s participation reflects SERC, IIIT Hyderabad’s ongoing research and engagement in the emerging area of Agentic AI and dependable AI-enabled software systems.

August 2026