This course is available on the MSc in Applied Social Data Science, MSc in Behavioural Science and MSc in Social Research Methods. This course is freely available as an outside option to students on other programmes where regulations permit. It does not require permission. This course uses controlled access as part of the course selection process. For information on controlled access courses, including eligibility, application processes, deadlines, and departmental contact details, please refer to the Controlled Access Courses webpage.

This course has a limited number of places (it is controlled access). Priority will be given to students on the MSc Applied Social Data Science and MSc Social Research Methods.

How to apply: For information on how to apply for a place on this course, including deadlines, please see this page.

This course is freely available as an outside option to students on other programmes where regulations permit. It does not require permission.

Assumed prior knowledge:

Introductory-level probability and statistics Experience programming for social science research (e.g., in Python or R)

Generative AI (particularly large language models, but also speech, image, and multimodal tools) is rapidly changing how social scientists work with text, data, and evidence. This course teaches students to use these tools effectively, critically, and ethically in their own research. It is organised around the problems researchers and practitioners actually face, such as reviewing literatures, coding open-ended survey responses, analysing policy documents, extracting structured data from text, simulating survey responses for pretesting, rather than around the technology itself. The course progresses from individual interactions with models, through evaluation and model selection, to multi-step workflows and applied research tasks.

Students begin with the tools most people encounter first, interactive systems such as Claude or ChatGPT, and learn how to structure prompts and manage tasks effectively. The course then moves to programmatic workflows, showing how models can be accessed through APIs and integrated into reproducible research pipelines. Students also learn how to navigate the model ecosystem, making informed choices about which models to use, how to access them, and what the trade-offs are around cost, privacy, and reproducibility.

A central theme is evaluation and reliability: how to validate model outputs, when they can be treated as research data, and how to maintain human judgement throughout. Responsible and reproducible AI use is emphasised throughout, with attention to data security, academic integrity, and transparency in reporting.

20 hours of lectures and 10 hours of seminars in the Winter Term.

This course has a reading week in Week 6 of Winter Term.

Problem sets.

Formative assessment consists of weekly exercises introduced during seminars and completed in students' own time. These exercises focus primarily on Python-based coding tasks and are designed to build the practical skills assessed in the summative components. Sample solutions are provided so that students can compare their own work and identify areas for improvement.

• Jurafsky, D., & Martin, J. H. (2024). Speech and Language Processing (3rd ed. draft).
• Bail, C. A. (2024). Can generative AI improve social science? Proceedings of the National Academy of Sciences, 121(21).
• Törnberg, P. (2024). Best practices for text annotation with large language models. Sociological Methods & Research.
• Argyle, L. P., et al. (2023). Out of one, many: Using language models to simulate human samples. Political Analysis, 31(3), 337–351.
• Spirling, A. (2023). Why open-source generative AI models are an ethical imperative for social science. Nature Computational Science, 3, 1014–1017.
• Bender, E. M., et al. (2021). On the dangers of stochastic parrots: Can language models be too big? Proceedings of FAccT 2021, 610–623.

Exam (80%), duration: 120 Minutes in the Spring exam period.

Project (20%).

The summative mark comprises two components: a take-home project (20%) due during WT, and a final examination (80%) in ST.