Artificial intelligence is becoming part of everyday life. We use it to search for information, generate text and make decisions, and lawyers and judges are beginning to use it too.
But could AI ever help interpret the law?
That is the question being explored by Morton Thornton, whose PhD, funded by the Alumni Postgraduate Research Scholarship, examines whether natural language processing, the technology that enables computers to analyse and generate human language, has a place in legal interpretation.
Judges regularly deal with words and phrases that are vague, open-ended or capable of more than one meaning. This is sometimes described as the “open texture” of the law. A word may appear straightforward until it is applied to a new situation, when its meaning becomes less certain.
Morton’s research asks whether Large Language Models (LLMs), which power many of today’s chatbots, could help judges interpret this kind of language.
LLMs are trained on vast quantities of written material. They analyse patterns in language and use those patterns to produce responses that can sound remarkably human. However, the way an LLM generates language is fundamentally different from the way people communicate and reason.
Why this matters
The pressure to use AI within the justice system is growing. In the future, some routine legal decisions may become automated. As AI becomes more common in legal settings, it may eventually be considered for more complex cases, including those heard by higher courts.
Before that happens, Morton believes we need to ask more than whether AI would be useful or efficient. We also need to understand whether it is capable of legal interpretation at all, and what its use would mean for responsibility, accountability and justice.
One insight emerging from his research is that, despite rapid advances in AI, the principles underlying the technology remain much the same as those behind early language models. Although chatbots can now produce language that seems remarkably human, their responses are generated through complex prediction rather than human reasoning.
“It’s easy to think of AI as being human-like in the way it processes information.”
As Morton works towards completing his research in 2027, he cannot stress enough his gratitude to the alumni community, whose generosity has made this academic journey possible:
” Without this scholarship, I doubt my PhD journey would have got off the ground in the first place. Everything I’ve done, and everything I hope to achieve, is thanks to this scholarship.”