AI systems are known for their confidence, but that doesn't always mean they're right. In a recent panel discussion recorded live at Qlik Connect, experts explored the phenomenon of AI hallucinations—where AI generates plausible but completely incorrect information.
Frederic Van Haren, CTO and Founder of HighFens, and Sidney Drill, Qlik Global Solutions Director for Data, Analytics, and AI, joined host Stephen Foskett to illustrate the issue with a real-world example: the "Paper Plane" cocktail. An AI system confidently produced a fabricated origin story for the drink, highlighting how even widely used AI can mislead.
The conversation delved into the underlying causes, including the cost and scalability challenges of AI agents, and the risks of relying on unreliable outputs. The panel emphasized that strong data foundations and governance are essential for building trustworthy AI in the enterprise. Without proper oversight, AI hallucinations can erode confidence and lead to costly mistakes.