Thomas Hughes introduced his 1857 novel Tom Brown’s School Days with the famous opener:
“The Browns have become illustrious by the pen of Thackeray and the pencil of Doyle…”
Hughes was referring, of course, to William Makepeace Thackeray and Richard Doyle, two prominent contributors to Punch. 169 years later, the Browns have once more been made illustrious, but this time with a little help from artificial intelligence and legal research methodology.
Therefore I’m pleased to share that a new article, co-authored with Ryan Brown, has been accepted for publication in the Journal of Legal Research Methodology.
‘The Reflexive Local AI (RLAI) Framework: a methodology for AI-assisted transcription in empirical legal research’ develops a framework for the responsible use of locally deployed artificial intelligence in empirical legal research.
The article grew out of my British Academy / Leverhulme-funded project, Pensions in a Warming World: An Empirical Study of the Opportunities and Challenges (2025-2027), for which I conducted 33 semi-structured interviews. Faced with the practical, methodological and data-protection challenges involved in transcribing and processing sensitive qualitative research data, we developed an approach using locally deployed AI tools rather than cloud-based AI services.
When I first designed the empirical study, transcription costs were £6,000 for 20 hours recordings. For a small research grant, that would have consumed a substantial proportion of the project’s available resources and significantly constrained the research.
The solution emerged somewhat unexpectedly. That evening, at a BBQ at my brother’s home, I mentioned the problem to Ryan. He suggested that I explore running AI transcription locally – an approach I had never previously considered and, in truth, knew very little about. That conversation ultimately provided the starting point for what became the RLAI Framework.
There is, of course, nothing new about researchers and writers confronting the cost of turning recorded or handwritten material into typescript. By coincidence, while browsing YouTube recently, I came across a 1964 interview with J.R.R. Tolkien (https://www.youtube.com/watch?v=z5PHvmf_–o). Discussing the production of his work, Tolkien recalled having to type it himself because he couldn’t afford the cost of professional typing:
“I typed that whole work out twice and lots of it, many times, on a bed in an attic. But then I couldn’t afford the um cost of the typing. There’s some mistakes too…”
More than sixty years later the technology is radically different, but the underlying problem struck me as remarkably familiar: what happens when the cost of outsourcing an essential stage of producing knowledge becomes a barrier to the work itself?
Local AI offered one answer to that problem, and more broadly is a means to democratise empirical study. But adopting it also raised a much more interesting set of questions about confidentiality, data protection, research integrity and the limitations of automated transcription.
Ryan brought his expertise as a cloud engineer to the technical dimensions of the framework, while my contribution draws upon its development and application within empirical legal research.
The article has now been formally accepted and will shortly appear in the Journal of Legal Research Methodology. I’ll add the full citation and DOI here once the published version is available. I’m delighted that an unexpected conversation over a BBQ has ultimately resulted in a new methodological framework and a published article!
Dr Lloyd Brown, 18 September 2026
© Lloyd Brown 2026. All rights reserved. Lloyd Brown asserts his rights to be identified as the author of this work.
Leave a comment