Upcoming Events
Talking to Machines Seminar Series
Friday, 24th April 2026
10.00 – 11.00
Online: Zoom link
Professor Xun Pang
Professor · School of International Studies
Peking University | PKU Analytics Lab for Global Risk Politics
Professor Pang’s research spans global risk politics, the geopolitics of critical raw materials, and the application of LLMs to social science. She is the author of From Cold Politics to Hot Politics (Peking University Press, 2026) and a forthcoming textbook on Large Language Models and Social Science Research. She has published in Political Analysis, International Organization, and Political Science Research & Methods, among others.
Yang Wu
Ph.D. Candidate
Institute of Automation, Chinese Academy of Sciences
Yang Wu’s research focuses on large language models, LLM reasoning, social simulation, and computational sociology.
How Can Synthetic Experiments Deliver Credible Causal Inference in Social Science?
The rapid diffusion of large language models (LLMs) has spurred growing interest in ‘synthetic experiments’, in which LLMs or LLM-driven agents simulate human subjects for causal inference. While such approaches promise scalability, cost efficiency, and experimental flexibility, fundamental methodological challenges — both theoretical and technical — must be addressed before they can serve as a credible causal engine.
Drawing on a pilot study that synthetically replicates the ‘hawkish bias’ experiment in foreign policy decision-making, this talk identifies key obstacles to credible causal inference in synthetic settings — including persona drift, ambiguous treatment assignment, underdeveloped benchmarking, and an unsettled research design — and discusses potential solutions from a learning perspective.
This session will cover:
- The rise of synthetic experiments using LLMs for causal inference in social science.
- Key methodological challenges: persona drift, treatment assignment, benchmarking, and sampling.
- How counterfactual data in fine-tuning promotes causal rather than correlational learning.
- Reasoning-oriented training for capturing intermediate causal mechanisms.
- Practical implications for designing credible AI-driven social science experiments.
Talking to Machines Synthetic Replication Games Workshop @ IMEBESS/EPSA
EPSA Prague 2026 Pre-Conference Event
Date: Wednesday, 8th July
Venue: Prague Congress Centre, Prague, Czech Republic
Format: Presentations + round table + hands-on programming workshop
Talking to Machines, Oxford is hosting a pre-IMEBESS/EPSA round table exploring the use of large language models as synthetic subjects in social science research.
