"Then there's a mistake... – How much sloppiness do we allow in AI?"

Errors are part of everyday life – but what happens when we start accepting them as “normal” just because they come from AI? In this episode of the AI Kantine, Buschi talks about an interview with Prof. Holger Hoos, who warns of a growing tolerance for errors in AI applications. Why generative AI is often unreliable, why we are starting to get used to “good enough,” and what the difference is between creative and dangerous errors – that’s what this critical, yet practical lunch break episode is all about.