Skip to content

Research project

Accelerating Social Science Practice in the Age of AI

The internet, and now AI, have accelerated changes in many aspects of human behaviour. This poses both opportunities and challenges for social science. Although the past decade has seen an enthusiastic adoption of Big Data and computational social science approaches, the institutions of social science have not sufficiently adapted to the current techno-social reality.

This project has received funding via the EUI Research Council call 2024.

We are currently witnessing one of the most transformative technological revolutions in history, driven by advancements in information technology and artificial intelligence (AI). Yet, the systems that underpin the production and legitimation of knowledge in academic institutions remain rooted in the technology of the printed word. Despite radical changes introduced by digital innovations, academia continues to rely on "papers" to encode and disseminate knowledge—a practice akin to using a supercomputer to read a phone book. While efficient, this approach underutilizes the transformative potential of new technologies.

This project explores how the sociology of social science and the status-driven lifecycle of researchers shape our adherence to traditional academic practices. It advocates for a reorganization of social scientific practices in response to the internet and AI revolution, aiming to fully harness the potential of these technologies.

A key focus is the development of theoretical frameworks for what Duncan Watts terms solution-oriented social science. While the burgeoning field of computational social science has made strides in this direction, much of the discourse around AI in social science oscillates between abstract enthusiasm and a conservative defence of the status quo. This project seeks to move beyond these extremes by:

  1. Quantifying Current Practices: Developing empirical insights into how social science is presently adapting to AI.
  2. Modelling Future Possibilities: Constructing computational models to envision how social science methodologies can evolve in an age of AI.

The project also contributes to meta-science, the study of how scientific knowledge is produced and validated. By expanding the scope of research into methodology to include meta-scientific critique, it aims to improve the statistical robustness and societal relevance of social scientific knowledge. For instance, innovations like results-blind publication formats address systemic issues such as the “file-drawer problem” (where studies with null results are underreported), improving the reliability of published research.

Ultimately, this project seeks to address two central questions:

  1. How is social science currently adapting to the challenges and opportunities posed by AI?
  2. How should it adapt—how should social science be organized to maximize its potential for innovation and societal impact?

Through this work, we aim to reimagine the practices of academic knowledge production, ensuring they are well-suited to the digital and AI-driven era.