Funded projects
Selected funded projects (click on the projects to find out more).
The project explores how information retrieval based on large language models can support access to legal information for lay users facing everyday legal problems. While legal technology is expanding rapidly, many automated legal services remain oriented towards commercial legal practice and profitable areas of law, leaving high-volume areas involving ordinary court users comparatively underserved. Focusing on private family law, the project develops and tests a GraphRAG-based approach for retrieving and explaining relevant legal information from complex legislative texts. Drawing on UK legislation converted from structured XML into graph-based knowledge representations, the project illustrates how knowledge engineering rooted in knowledge graph approaches, prompt engineering, and GraphRAG can help identify relevant legal information, structure it around legal needs, and support more transparent retrieval suitable for lay court users. By comparing the same user scenarios across different graph variants and work flows, the project investigates whether richer legal structure improves retrieval relevance, contextual completeness, traceability, and answer quality. The objective of the project is to evaluate to what extent carefully designed and tested LLM-based legal tools can contribute to revolutionising and redefining open justice principles, particularly by making complex legal materials more accessible, explainable, and responsive to the needs of users across a wider range of legal matters.
The research study initiated a new collaboration with police forces to explore the challenges and opportunities of current Domestic Abuse (DA) risk assessment tools used in the combined jurisdiction of England and Wales. The first stage of the study involved interviews with DA desk duty officers and review and management staff and discursive textual ethnography of the risk assessment tools. The results illustrated that risk assessments conducted via video link present a unique opportunity to enhance victims’ perceptions of procedural justice, though other stages of the reporting and investigative journey and other methods of conducing risk assessments require further investigation and training. The second stage of the project aims to understand policing practices across all individual stages of investigations separately and as part of the case progression. The overall objective of the project is to deliver a quantitively and qualitatively informed evidence base, drawing on the combination of case-level and force-level data and mixed methods approaches to understand the end-to-end systemic, operational and procedural factors impacting DA responses, case progression and outcomes.
I led the project ‘Language of DIY Justice: Communication Practices & Processes’ funded by AHRC Standard Grant scheme.
The project drew on quantitative and qualitative mixed methods approaches to investigate the following aspects: communication as part of procedural justice and social justice; legal-lay discourses across different stages of legal proceedings; elicitation strategies used in court forms, court users’ experiences when using digitised court services, and narrativisation procedures across different types of civil proceedings. The wide scope of the project led to the discovery of issues with the language of court forms, communication during online courts, or narrativisation procedures in legal settings. For instance, the research conducted for the project illustrated that civil and family procedure rules routinely mute participants’ voices, mismanage language data, complicate the judiciary’s role and unnecessarily delay the presentation of parties’ narratives. The project results are presented across five journal articles and the forthcoming Cambridge Element.
Focusing on adversarial legal settings, the Cambridge Element titled Legal-Lay Discourse of County and Family Courts (2024, in print) explores discursive practices in court proceedings which often involve unrepresented parties – private family proceedings and small claims cases. Such proceedings present the main caseload of county and family courts, but pose immense challenges when it comes to legal-lay communication. Drawing on court observations, alongside textual and interview data, the Element pursues three aims: (1) developing the methodological and theoretical framework for exploring discursive practices in legal settings; (2) establishing the link between legal-lay discourse and procedural justice; (3) presenting and contextualising linguistic phenomena as an inherent part of court research and practice. The Element illustrates how linguistic input can contribute to procedural changes and court reforms across different adversarial and non-adversarial legal settings. The exploration of discursive practices embedded in court processes and procedures consolidates and advances the existing court research conducted within the fields of socio-legal studies and forensic linguistics.
I was a co-investigator on the project ‘TRAC: COVID – Trust And Communication: a Coronavirus Online Visual Dashboard’ funded by AHRC UKRI COVID-19 Research and Innovation.
The project collected over 80 million tweets and applied corpus linguistics methods to visualise the conglomerated data and create a dashboard traccovid.com. As part of the project, I led a case study on the Government management of the COVID-19 communication and public perception of the pandemic. The case study and the research findings were accepted as evidence for the report on Initial lessons from the government’s response to the COVID-19 pandemic released by the House of Commons, Committee of Public Accounts and the National Audit Office. The case study and the project showcase the role public humanities play in enhancing policy-making and measuring the public trust in government bodies.
I led the project ‘Changing landscape of access to justice: Linguistic and socio-legal analysis of online forums for litigants in person’ funded by BA/Leverhulme Small Grants.
The project explored a grey area of the provision of legal advice online by analysing the quality and type of advice provided by McKenzie Friends, unregulated lay advisers who offer legal advice to court users for free or for a fee. The project showed that online forums and social media led by McKenzie Friends create a communicative environment in which the linguistic framing of the advice jeopardises its potential usefulness and creates a false sense of injustice, arguably misleading and potentially financially exploiting vulnerable users. The project also led to the discovery of common misconceptions among court users and the wider public, contributing to the research agenda of legal public education.
The project featured in professional publications and practitioner print, such as Legal Futures, The Law Society Gazette, The Times, Today’s Family Lawyer, Lawyer Monthly, Ministry of Justice Newsletter, which built on the project findings to promote the regulation of the lay advisers’ services and push for the need to provide clarity for consumers. As the only quantitative and qualitative empirical study of McKenzie Friends’ out-of-court services, the project provides a strong evidence base for reviewing the regulatory framework of the legal advice provision.
I led the project ‘Pro se language use’ funded by EU Marie Curie Intra-European Fellowship scheme for Career Development Grant.
The project examined a corpus of trial transcripts from cases with litigants in person tried in England and Wales, USA and Canada. The focus was on communication challenges experienced by unrepresented litigants when conducting witness examination, delivering opening and closing arguments and discussing procedural matters with judges and legal representatives for the opposing party.