Does AI just speak, or does it already think? On March 24, 2026, our principal investigators Prof. Birte Platow and Prof. Martin Potthast joined the podcast Dienstags direkt to discuss this question with Prof. Stephan Buchhester (Institut für Verhaltensökonomie Leipzig). The episode was moderated by Thomas Lopau (MDR Sachsen). You can listen to the full podcast episode here.

Since the release of ChatGPT, language models have become part of our everyday lives. Even non-experts began interacting with this new technology and were just as fascinated by it as the experts. Numerous possibilities opened up: summarizing texts, drafting well-thought-out emails, or even illustrating bedtime stories for one’s own children. But concerns soon began to arise. Will AI take our jobs away? Why do we even need to learn new skills if AI can handle difficult tasks?

In this episode of Dienstags direkt, Prof. Birte Platow, Prof. Martin Potthast, and Prof. Stephan Buchhester discuss why humans are and will stay important in technological progress. Not only does the quality of a chatbot answer depend on the quality of the prompt, but language models can also hallucinate. Such hallucinations are plausible-soundig but factually incorrect answers. Therefore, a review of generated answers remains absolutely necessary. AI can’t replace human judgment, but it can drastically increase the speed in which we can solve tasks. This is especially useful for tedious routine tasks that we may encounter in our personal lives or at work.

At the same time, we need to protect ourselves. Humans need a sense of self-efficacy – even if that means solving tasks ourselves without the help of technology. Of course we could quickly delegate tasks – such as writing a birthday card – to AI, but it is our creativity, our emotions, and our personal thoughts that give meaning to these actions.