LaclauGPT
LLM-assisted computational discourse analysis for social and political research.
LaclauGPT is an open social-science research framework for studying large textual and multimodal corpora with computational methods and LLM assistance. It keeps interpretation traceable to source evidence, uncertainty, provenance, and human researcher review.
Research status
Draft research plan. The current Phase 1 paper is a working research-plan manuscript. It has not been formally published or peer reviewed, and its arguments, methods, scope, and wording may change as the research develops.
About and authorship
LaclauGPT has been under development by Tomi Toivio since 2023.
Contact: tomi.toivio@helsinki.fi · GitHub profile
The project was originally created for the EP24 research project in connection with the Helsinki Hub on Emotions, Populism and Polarisation (HEPP) and the Helsinki Institute for Social Sciences and Humanities (HSSH). The current version is guided primarily by the AI26 research project, while also supporting other research projects and methodological development.
Project background and funding
LaclauGPT grew out of research-software work at the Helsinki Hub on Emotions, Populism and Polarisation (HEPPsinki), University of Helsinki, and has been developed in connection with the CO3, PLEDGE, and ENDURE research projects.
- CO3 — Continuous Construction of Resilient Social Contracts Through Societal Transformations.
- PLEDGE — Politics of Grievance and Democratic Governance.
- ENDURE — Inequalities, Community Resilience and New Governance Modalities in a Post-Pandemic World.
Funding. CO3 and PLEDGE have been funded in the Horizon Europe framework of the European Union (2024–27). ENDURE was a Trans-Atlantic Platform funded consortium; the work at the University of Helsinki was funded by the Research Council of Finland (2022–25).
Active project
The current LaclauGPT architecture consists of four active repositories:
| Repository | Role |
|---|---|
| LaclauGPT | Project hub: scientific paper, theory, architecture, shared contracts, roadmap, and complete-system documentation |
| LaclauGPT Data Collection | Collect: acquisition, capture, normalization, provenance, scheduling, and storage adapters |
| LaclauGPT Data Analysis | Analyze: NLP, embeddings, LLM-assisted discourse analysis, statistics, uncertainty, and evidence-linked analytical proposals |
| LaclauGPT Data Visualization | Visualize and review: dashboards, timelines, maps, graphs, corpus exploration, and human researcher review |
The research pipeline is:
Data Collection → Data Analysis → Data Visualization
↑
LaclauGPT hub
theory · paper · contracts
Research
The current flagship research programme is LaclauGPT: Ideological contestation over AI. The framework is grounded in discourse theory associated with Ernesto Laclau and Chantal Mouffe and is designed for human-in-the-loop computational social-science research.
- Scientific paper
- Theory and methodology
- Roadmap
- Canonical data contract
- Plugin platform architecture
Historical research archive
Two earlier repositories are preserved as historical research artifacts. They document the lineage of LaclauGPT but are not part of the current software architecture.
| Repository | Historical role |
|---|---|
| LaclauGPT Multimodal Analysis | Legacy EP2024 multimodal analysis pipeline and research documentation |
| LaclauGPT TikTok Scraper | Legacy EP2024 TikTok collection tooling and research documentation |
For current development, start with the four repositories in Active project above.
Research-use boundary
Human-in-the-loop academic research only. Machine-generated interpretations are preliminary analytical proposals. Source evidence and interpretations must be verified by a human researcher before they are reported or cited as research conclusions.
Source code
All active public development is organized under the LaclauGPT GitHub project.