GenieAI Legal AI Lab
Our research.
Explore AI research we've completed alongside advisors Professor Jun Wang, Lord Neuberger, and other industry titans.
What our research covers
Studying generative AI and the legal industry since 2017.
We have been studying the effects of generative AI on the legal industry since our company was founded in 2017 - long before the wider market caught onto the technology's potential. Our work spans peer-reviewed academic research published at top NLP conferences, a multi-year UKRI-funded grant project with Barclays and Oxford University, the most rigorous public benchmarks of legal AI tools available today, and a 67-page whitepaper synthesising eight years of frontline learning across in-house, Biglaw, and small-firm contexts. Here's a brief summary of what we've worked on - and where we think the industry is going next.
De-identification of Contracts
In collaboration with Barclays, Withers, the University of Oxford, and Imperial College London.
UKRI awarded a £1.5M grant to GenieAI's consortium - Barclays, Withers, the University of Oxford, and Imperial College London - to answer two of the biggest challenges to the adoption of AI in legal services:
- access to confidential data;
- understanding the decisions machine learning models make.
The project aimed to answer these challenges with the combined capabilities of:
- GenieAI's Legal AI
- Oxford University's research into robust recurrent neural networks to explain the accuracy of machine-learning decisions
- Imperial College's research into the de-identification of legal contracts to anonymise confidential data while retaining critical information
- GenieAI
- Barclays
- Withers
- University of Oxford
- Imperial College London
The State of Legal AI
Download our full 67-page whitepaper covering 92 topics across in-house, Biglaw, and small-firm legal AI - published on December 12th 2024.
Comparisons of Legal AI tools
We researched 80+ legaltech tools claiming to offer AI services, and compare them side-by-side.
The result is a growing library of head-to-head comparisons - dozens of published Genie-vs-competitor breakdowns covering price, accuracy, jurisdictional coverage, and workflow fit - plus rigorous benchmarking studies measuring how Genie performs against ChatGPT, Claude, and other LLMs on real legal tasks.
Much more to come
Four years before the world had its ChatGPT moment, our co-founder and CEO Rafie Faruq was writing his thesis on generative AI. A lot has changed since - and we have unpublished research on effective prompting of legal AI, anonymisation, document transformation, AI negotiation, and much more. We apply first-principles thinking within our senior machine-learning team here, tied to our relentless pursuit of delivering value to our customers.