AI and society: Social physics, or the interpretative turn?

An interdisciplinary symposium at the University of Warwick

24 June 2025

Register to attend using the sign-up form below!

There is increasing overlap and exchange between computational, mathematical and social sciences in an age defined by AI. More and more social researchers take up computational methods, while computational scientists analyse intrinsically social phenomena. Furthermore, during the last few decades new areas of research have emerged at the interface between social sciences and computational sciences, such as computational social science and digital sociology. This symposium will explore the overlaps and exchanges between these various fields with a special focus on foundational concepts and methodological principles.

Topic

In the trading zone (Galison, 2010) between social and computational science, new possibilities and fundamental questions arise. Some argue that the combination of AI and big data will make possible the discovery of social laws (Lazer, 2023). Others claim that machine learning will unleash a new era of sociological theory and will finally enable the development of a qualitative computational science of society (Borch & Pardo-Guerra, 2024).

In the 19th century, statistics and sociology found joint origins in the new methodology of social physics. In the wake of social media, attempts were made to reinvent this tradition, with mixed success (Watts, 2016). More recently, "AI" has been welcomed as an important opportunity for the renewal of interpretative social science (Friese, 2023; Uitermark and Tonberg, 2024). Can the wide uptake of machine-learning based techniques across fields provide occasions to bring these debates and initiatives to maturity? Can we operationalise core questions and methodologies by combining social and computational theory and methods in new ways?

Themes

To explore these broad questions through interdisciplinary exchange, we propose the following themes:

1. Social laws and social dynamics: determinism, scale, contingency

Can machine-learning methodologies bring social laws within the purview of an empirical computational social science? Is social life fundamentally dynamic as well as contingent?

2. Concepts of society: social systems, social agents, social life

What concepts of society and social life are and are not operationalisable within ML-based frameworks?

3. Population, environment and individuals: definitions, properties and fallacies

How do we define units of analysis in computational science and social science? What relation between data, analysis and context of application do they make possible?

Format

This one day symposium will bring together Warwick researchers, their research students and external speakers to explore opportunities for interdisciplinary exchange between computational and social sciences.

We plan to organise the discussion around the above three thematic areas. In addition, introductory talks will bring together fields through a focus on:

- Principles & practice of Social Science/Sociology

- Principles and practice of Statistics/ML/AI/CSS

- The origins and intertwined history of Social Science/Sociology and Statistics/ML/AI/CSS

For each theme, we will invite senior and early career speakers from different disciplines to prepare contributions.

Organising team

Noortje Marres (Centre for Interdisciplinary Methodologies),

Theo DamoulasLink opens in a new window (Statistics/Computer Science),

Greta Timaite (Centre for Interdisciplinary Methodologies),

Federico Perlino (Statistics)

Yorgos Felekis (Computer Science)

Hosted by the University of Warwick’s Society and Culture SpotlightLink opens in a new window in collaboration with the Digital SpotlightLink opens in a new window.

The symposium programme will be sent to all registered participants in due course