Artificial Unintelligence - Why AI lies with a straight face
Ever caught a chatbot lying with confidence? This episode of Thats Science explains why AI gets things wrong and why some errors stay.
Have you ever been having a perfectly normal chat with your favorite AI large language model, only to realise it's been confidently lying to your face?
In this episode of That’s Science, we dive into the glitchy, often confusing world of AI hallucinations - those moments when a large language model confidently feeds you a lie with a straight face. You aren't alone if you’ve felt misled by a chatbot.
Joining us as Dr Wei Zing to break down why these digital delusions happen and explore the surprising reality that while developers have the power to eliminate these errors, they might just choose to keep them around.
Recommended Reading:
Credits
- Host: Tabby Taylor Buck (Freelance Content Creator)
- Guest: Dr Wei Xing (Lecturer in Mathematics at the University of Sheffield)
- Podcast and Video Editor: Sophie Cowell (Freelance Content Creator)
- Image: Created with Google Gemini
- Produced by: Shaun Oliver (Marketing and Communications Manager at the University of Sheffield)
- Produced by: Tabby Taylor Buck (Freelance Content Creator)
- Support from: David Heath (Digital Engagement Officer at the University of Sheffield)