Anna Jobin, Professor at the Interdisciplinary Human-IST Institute of the University of Fribourg, discusses AI hallucinations and challenges they pose in the French online publication Watson.

In an interview with Alexandre Cudré, she explains how AI-generated answers can appear plausible while actually being incorrect or inconsistent, and emphasizes the need to remain vigilant and verify information using other sources. She also discusses the responsibility of technology providers behind these models to make AI systems safer.

Read the full interview (in French): https://www.watson.ch/fr/suisse/technologie/958113244-l-ia-hallucine-et-c-est-important-de-savoir-comment-ca-marche