JICL publication: A generative AI-based legal advice tool for small businesses in distress

· Thomas Wood
JICL publication: A generative AI-based legal advice tool for small businesses in distress

We are pleased to announce the publication of our paper A generative AI-based legal advice tool for small businesses in distress, in collaboration with an interdisciplinary team based in the UK and Hungary.

This paper describes the development and evaluation of the Insolvency Bot, a legal chatbot designed to provide reliable advice on corporate insolvency in England and Wales for small business owners.

We used Retrieval Augmented Generation (RAG) to enhance large language models with a curated knowledge base of 6,000 legal texts, including statutes, HMRC forms, and case law.

Thomas Wood, the director of Fast Data Science, was responsible for building the system in Python and implementing the machine learning models and vector embeddings that allow the bot to retrieve relevant legal information. Fast Data Science hosts the live version of the tool on our website at https://fastdatascience.com/insolvency

Abstract of the paper

We developed and tested the performance of a retrieval augmented generation (RAG) system for answering legal queries related to corporate insolvency in England and Wales. The Insolvency Bot relies on open-source legal information and HMRC forms to provide sound responses to a user’s query focusing on insolvency matters regulated by English law. We evaluated our bot head-to-head on an unseen test set against the unmodified versions of large language models (LLMs) gpt-3.5-turbo, gpt-4, or gpt-4o with a mark scheme similar to those used in examinations in law schools. The Insolvency Bot outperformed each unmodified LLM (p = 0.05%). An additional user experience survey suggested the need for creating two versions of the bot, one for lay people who expect practical and actionable advice and another for professionals with the relevant legal authorities. Our legal chatbot demonstrates the benefits of combining a generative AI system with a trusted knowledge base and shows future promise to cover cross-jurisdictional and insolvency-related queries and could be further improved in its technical architecture.

Citation

More information: https://pure.royalholloway.ac.uk/en/publications/a-generative-ai-based-legal-advice-tool-for-small-businesses-in-d/

Find Top NLP Talent!

Looking for experts in Natural Language Processing? Post your job openings with us and find your ideal candidate today!

Post a Job

Should AI companies be allowed to mine creative content to train models?
Legal ai

Should AI companies be allowed to mine creative content to train models?

The effects of AI companies training on creative works When AI companies train models on text, video and images from creative industries, the end product is an AI model which can create near-human quality visual designs or copywriting. Many artists have argued the AI companies are exploiting creatives by profiting from their work, while the original creators do not receive any compensation. By flooding the market with low-cost content, AI has driven down rates for human artists, which is threatening creative careers.

Are we in an AI bubble?
Generative ai

Are we in an AI bubble?

None of the content of this article constitutes investment advice. Few would dispute that the global economy is currently in an AI boom. However, opinions are divided over whether this boom also constitutes a bubble.

Launching Harmony Meta
Ai in research

Launching Harmony Meta

We are excited to introduce the new Harmony Meta platform, which we have developed over the past year. Harmony Meta connects many of the existing study catalogues and registers.

What we can do for you

Transform Unstructured Data into Actionable Insights

Contact us