Social and ethical perspectives on Agentic AI
What happens when AI systems no longer just provide answers, but increasingly act on their own? This question was at the center of The Agentic Enterprise – Swiss Executive Forum, held in Zurich on September 10, 2026. The event brought together perspectives from academia, business, and technology to discuss the opportunities and challenges of Agentic AI.
Representing the University of Zurich,Prof. Dr. Anikó Hannák from the Department of Informatics and law professor Prof. Dr. Florent Thouvenin contributed the academic perspective on the societal, ethical, and legal implications of increasingly autonomous AI systems.
In her contribution, Anikó drew on insights from research on socio-technical systems to examine which long-standing risks and considerations become particularly relevant as companies move towards Agentic AI. These include biases inherited from training data, unequal access and performance across different user groups, and the potential homogenization of knowledge and expertise. She also addressed questions of trust and accountability: as employees increasingly delegate tasks to AI systems, their role may shift from performing tasks themselves to monitoring and evaluating AI-generated outcomes.
For organizations adopting Agentic AI, this means looking beyond productivity gains alone. Anikó emphasized the importance of understanding what AI systems were designed for, equipping employees with the skills to critically evaluate their outputs, and considering the trade-offs that come with greater reliance on AI.
The forum concluded with a panel discussion bringing together perspectives from research and industry on how organizations can make use of increasingly autonomous AI while addressing questions of control, responsibility, and governance.