Reflection on National Conference "Autonomy in a Digital Era"

Reflecting on the National Conference “Autonomy in a Digital Era”

On 30 January 2026, Utrecht University hosted the national conference "Autonomy in a Digital Era" at the Railway Museum in Utrecht. Organized jointly by the focus areas Applied Data Science, Governing the Digital Society, and Human-Centered Artificial Intelligence, the conference brought together researchers from across disciplines to reflect on how digital technologies are reshaping the meaning and conditions of autonomy. The presentations of the speakers are available on the conference website. 

The programme combined perspectives from artificial intelligence, sociology, economics, philosophy, media studies, and biomedical science. Discussions ranged from adaptive AI agents and human–AI collaboration to platform economies, digital sovereignty, environmental consequences of large-scale AI infrastructures, and medical applications of machine learning.

The interdisciplinary set-up proved especially valuable. Researchers working on the technical optimization of AI systems were confronted with questions from the audience concerning the societal implications and potential risks of their work. Conversely, scholars approaching AI from political and social perspectives — often focusing on the dangers of uncritical adoption and the risk of AI reinforcing existing power asymmetries — were challenged on the technical assumptions underlying their analyses. Rather than remaining parallel conversations, these exchanges created a productive dialogue in which empirical, technical, and normative perspectives continuously informed one another.

These differences became particularly visible during the concluding panel discussion, where speakers offered contrasting answers to the question whether large language models can be considered autonomous. One view characterized LLMs primarily as powerful tools that amplify the capacities of actors who already possess social and economic power. Another emphasized their potential to enhance and scaffold the autonomy of users. The disagreement revealed that answers to such questions depend not only on normative evaluation but also on what is technically possible in modelling autonomy and on whether increasingly sophisticated AI systems can meaningfully be said to embody forms of agency. In this discussion, political analysis, ethics, theoretical AI, and philosophy came together, illustrating the value of sustained interdisciplinary engagement.

With a large and engaged audience of researchers, professionals, and students, the conference succeeded in fostering new conversations and collaborations. Rather than arriving at a single definition of autonomy, the event demonstrated the productivity of approaching the concept from multiple disciplinary angles — an approach that will continue to guide future research initiatives at Utrecht University.

Jan Broersen and Mehdi Dastani
Chairs Focus area Human-centered Artificial Intelligence