Can ChatGPT really understand language?

13 May 2026
Event
As artificial intelligence becomes increasingly embedded in daily life, questions about how AI systems process and generate language are attracting growing public attention. Can systems like ChatGPT actually understand what they are saying, or are they simply extremely sophisticated prediction machines?

These questions will take centre stage on 18 May during a public symposium in Amsterdam organised by the Genootschap ter bevordering van Natuur-, Genees- en Heelkunde (GNGH) and the Koninklijk Wiskundig Genootschap (KWG). The event will bring together researchers and the public to discuss how advances in artificial intelligence are reshaping ideas about language, cognition, and human intelligence.
Among the featured speakers is Micha Heilbron from our Institute.
 

Understanding language in humans and machines

Micha Heilbron studies how humans process and predict language in real time, and how these mechanisms compare to modern AI systems. His research combines neuroscience, psycholinguistics, and machine learning to investigate how both brains and language models anticipate upcoming words and meanings during communication.

Large language models such as ChatGPT can produce remarkably fluent and convincing text. But whether these systems truly “understand” language remains an open scientific question. By comparing the predictive mechanisms used by humans and AI systems, Heilbron’s work helps reveal both the strengths and the limitations of current artificial intelligence.

His research also contributes to a broader effort within cognitive science to better understand human language processing itself. Studying AI systems alongside the human brain offers researchers new ways to investigate how meaning, prediction, and communication emerge.
 

AI and society

The symposium comes at a time when generative AI is rapidly influencing education, communication, creative industries, and everyday work. Public discussions about these technologies increasingly touch on questions that have long been central to psycholinguistics and cognitive neuroscience: What does it mean to understand language? How do humans predict meaning during conversation? And can artificial systems ever achieve similar forms of understanding?

The event aims to make these discussions accessible to a broad audience by bringing together experts from different disciplines to reflect on the future of AI and human cognition.
 

Event information:

OBA Oosterdok, Amsterdam
18 May 2026
Free admission (registration required)
Publiekssymposium in samenwerking met het Koninklijk Wiskundig Genootschap, 18 mei 2026 – Genootschap ter bevordering van Natuur-, Genees- en Heelkunde

Attendance is free, but places are limited, so early registration is recommended.

Share this page