Connect
About
Jon Cardoso-Silva is an Associate Professor (Education) in the Department of Methodology. He teaches data science and AI to undergraduates and on executive education programmes, and he studies how generative AI tools affect student learning.
Jon is one of ten LSE AI and Education Fellows [1] (2025-2027), a programme that gives academics time and resources to reimagine how AI can transform teaching, learning, and assessment. Through the fellowship, Jon built an AI course tutor that scaffolds learning by being Socratic where appropriate but also giving direct help when needed, keeping pace with what the course has covered. The tutor has been deployed in two departments, with academic papers on the design and findings in progress. A model-agnostic guide for other staff is being published online soon, and he is exploring how to make the tutor more proactive to create a learning environment where AI supports thinking and engagement and avoids the cognitive offloading that generic chatbots often encourage. The fellowship grew out of GENIAL [2], a project he led with colleagues from several LSE departments where they collected chat logs, git histories, and coursework submissions to study how students use AI during coursework and whether those patterns affect what they learn.
Before joining Methodology in September 2026, Jon spent four years at the LSE Data Science Institute, where he led the teaching and development of two undergraduate courses (DS105 and DS205), acted as the institute's de facto deputy head of education, and represented the DSI at LSE teaching and AI fora. He received his PhD in Computer Science from King's College London in 2018, with research on optimisation, network science, and machine learning. He has prior industry experience as a software developer at Indra and as a lead data scientist at Data Science Brigade, a Brazilian consultancy.
Key expertise: AI in education, data science, mixed methods
[1] https://info.lse.ac.uk/staff/ESE/AI-Fellows-Jon-Cardoso-Silva
[2] http://lse-dsi.github.io/genial
Research
Research interests:
Jon studies the impact of generative AI on education, and in particular what students actually learn when they can hand the hard steps to an AI tool. He is interested in how usage patterns develop over a course, what those patterns tell us about how students are thinking, and what this means for how we design courses and assessments. His empirical approach combines qualitative coding of chat logs with trace analysis of version-control histories to build a detailed picture of how student work and AI use co-evolve. He developed the GENIAL framework for mapping these interactions onto experiential learning theory.
Jon also has a background in mathematical optimisation, network science, and machine learning, with publications on drug discovery modelling, epidemic forecasting, multi-omics network embedding, and latent network models for noisy social network data.
Publications:
Cardoso-Silva, J., Sallai, D., Kearney, C., Panero, F. and Barreto, M. (2025). Mapping Student-GenAI Interactions onto Experiential Learning: The GENIAL Framework. SSRN preprint.
Da Costa Avelar, P.H., Cardoso-Silva, J., Wu, M. and Tsoka, S. (2026). SCONE: A Subset-Contrastive Method for Multi-Omics Network Embedding. Briefings in Bioinformatics, 27(4), bbag381.
Sallai, D., Cardoso-Silva, J., Barreto, M.E., Panero, F., Berrada, G. and Luxmoore, S. (2024). Approach Generative AI Tools Proactively or Risk Bypassing the Learning Process in Higher Education. LSE Public Policy Review, 3(3), p. 7.
Li, Y., Cardoso-Silva, J., Kelly, J.M., Delves, M.J., Furnham, N., Papageorgiou, L.G. and Tsoka, S. (2024). Optimisation-based modelling for explainable lead discovery in malaria. Artificial Intelligence in Medicine, 147, 102700.
De Bacco, C., Contisciani, M., Cardoso-Silva, J. et al. (2023). Latent network models to account for noisy, multiply reported social network data. Journal of the Royal Statistical Society Series A, qnac004.
Cardoso-Silva, J., Papadatos, G., Papageorgiou, L.G. and Tsoka, S. (2019). Optimal Piecewise Linear Regression Algorithm for QSAR Modelling. Molecular Informatics, 38(3), 1800028.
Cardoso-Silva, J., Papageorgiou, L.G. and Tsoka, S. (2019). Network-based piecewise linear regression for QSAR modelling. Journal of Computer-Aided Molecular Design, 33(9), 831-844.