Use AI Hallucinations to Stimulate Your Critical Thinking

In this provocative and mind-expanding session from the Artificial Unintelligence Conference 2025, Marco Wobben challenges the common assumption that AI hallucinations are simply defects to be eliminated. Instead, he argues they can serve as powerful catalysts for creativity, inquiry, and deeper understanding of both human and machine cognition.

Marco explores how these “incorrect” outputs reveal the epistemic boundaries of AI systems — exposing where context, ambiguity, and missing information force humans to re-engage their own critical reasoning. Rather than treating AI hallucinations as failures, he shows how they can become tools for brainstorming, stress-testing ideas, and sharpening our skepticism in an age of increasingly fluent machines.

Key themes include:

Reframing AI hallucinations as cognitive and creative opportunities

Using generative errors to interrogate truth, context, and ambiguity

Developing productive skepticism as a form of modern AI literacy

Understanding the limits of machine reasoning versus human judgment

How human–AI co-evolution will shape the future of critical thinking