The Inaugural Large Language Models (LLM) x Law Hackathon

Written by The Organising Team: Zarja Hude, Annika Iyengar, Mahika Popli, Ryan Jackloski, Tomas Paulik, and Lara Kodrin


On 23rd June 2024, the King's Entrepreneurship Lab hosted its inaugural LLM X Law Hackathon, inspired by the well-established CodeX initiative at Stanford University. This event brought together a diverse group of students, academics, and industry experts from various backgrounds – law, tech, and many fields in between! It generated considerable energy and enthusiasm, surpassing all expectations. Here's a glimpse of what inspired the event, and what happened during those exciting 15 hours.

The Motivation

Why a hackathon?

Hackathons are intense, collaborative events where participants from diverse backgrounds create software solutions within a short timeframe to compete for prizes. These events typically occur in person and are led by industry experts. The focused environment enables participants to rapidly develop proofs of concept or working prototypes of potentially marketable solutions. The blend of student innovation and cutting-edge technology knowledge, augmented by industry experience, pushes the boundaries of what's possible.

Why LLM x Law?

The legal sector has been traditionally slow in adopting technology due to language barriers, complex legal data, and the high stakes of the industry. The emergence of Large Language Models (LLMs), however, is beginning to bridge the language barriers and spark significant excitement in the sector.

Other emerging technologies, like various forms of Machine Learning (ML), Graph Neural Networks (GNNs), Retriever-Augmented Generation (RAG), Chains of Thought (CoT), Trees of Thought (ToT), and more, also possess considerable potential for legal applications. This burgeoning field of research provides exciting opportunities for building approaches to complex legal challenges.

The potential of using the hackathon as a strategy to energise the link between LLM and Law has been recognised at CodeX, the Stanford Centre for Legal Informatics. For nearly two decades, they have conducted research at the intersection of law and computer science. One of their recent initiatives, the LLM x Law Hackathon, encourages the use of LLMs to create actionable solutions for real-world legal problems.

Why King’s College, Cambridge on 23rd June ?

The University of Cambridge, with its long-standing and well-established tradition of leading academic endeavour and nurturing original thinking, was an excellent venue to continue to explore the intersection of law and technology.

King's College is renowned as the home of Alan Turing, often referred to as the 'father of AI'. An innovator of his time and a pioneer of computer science, Turing’s legacy is enduring. In fact, the 'Turing Test', a criterion to determine whether a machine can exhibit behaviour indistinguishable from that of a human, is a phrase that permeates the consideration of LLMs and their future. It is perhaps less known that Turing was born on 23rd June.

The meaning, potential and timing of the hackathon at King’s thus holds particular significance and celebrates the home that King’s has provided for world-changers.

It felt that the inaugural LLM x Law Hackathon at King’s College on 23rd June is meant to be!

The Supporters

The Hackathon gained interest from leading firms in law, tech, and LegalTech, as well as from the public sector. This highlighted not only an enthusiasm for LegalTech but also an exploratory and collaborative spirit in the field.

The platinum sponsors of the event were Google Cloud, representing the tech sector, and Clifford Chance, representing the legal sector.

The golden tier of sponsors included:

•	Law firms: Linklaters, Simmons & Simmons, Brown Rudnick LLP, Pinsent Masons
•	Tech firms: Amazon Web Services (AWS) and Groq
•	Consultancies: D2 Legal Technology and Simplexico
•	LegalTech firms: Jylo (recently spun out from Travers Smith) and Robin AI

In addition, Mistral AI, BMW Group, A&O Shearman and Addleshaw Goddard contributed non-monetary prizes.

The event also received support from non-commercial entities, including Legal Tech Colab, a LegalTech accelerator, Publications Office of the European Union (EU), the Bar Standards Board, LawtechUK, and the European Central Bank.

The Participants

The Hackathon attracted a diverse group of students from multiple fields including STEM, law, business, linguistics and history.

A total of 150 participants from 12 countries travelled to Cambridge to participate. Students represented 33 universities.

The Day

The activities started on the evening of the 22nd of June at the "AI as a Canvas" garden exhibition, hosted by Simmons & Simmons. This event served as an introduction to the Hackathon, allowing participants to connect with sponsors and other participants while enjoying the sunshine and art in the King’s College gardens.

The Hackathon began early the next morning with the opening ceremony. Kimberley Kankam-Boadu, Legal Operations Professional at Google DeepMind, gave an inspiring keynote speech on the potential for LLMs to disrupt the legal industry. Participants then heard welcoming remarks from Kamiar Mohaddes, Co-Founder and Director of the King’s Entrepreneurship Lab and Faculty at the Cambridge Judge Business School, and Jay Mandal, a Stanford CodeX Fellow.

A significant feature of this Hackathon was the provision of problem statements by our co-organising institutions, sponsors, and partners. This structure allowed teams to maximise their time by tackling a predefined problem instead of identifying a problem from scratch. The most popular problem statements, requesting a method to extract information from contracts, were provided by Linklaters and Google Cloud.

For the next 8 hours, participants ideated and developed solutions to address the problem statements. Throughout the work period, representatives from our co-organising institutions, sponsors, and partners delivered 15 Lightning Talks in the historic Keynes Lecture Theatre. Topics included “Using LLMs in Legal Dispute Resolution” and “What is Stopping Clients from Adopting AI?”.

After the work period, participants submitted their solutions for judging. Each team prepared a short pitch to a different panel consisting of representatives from academia, law, and tech. Each panel selected 2-3 teams to advance to the Grand Finale.

Between the first round of judging and the Grand Finale, participants had the opportunity to attend Singing on the River, a concert featuring the world-famous King's Men performing on the River Cam (which was absolutely magical!).

After the concert, the top 12 teams competed in the Grand Finale. The judging panel comprised experts from our platinum sponsors and the co-organising institutions:

● Megan Ma from Stanford CodeX,
● Kamiar Mohaddes from King’s Entrepreneurship Lab,
● Hariprasad Radhakrishnan from Google Cloud,
● Anthony Vigneron from Clifford Chance, and
● Bernhard Waltl from the Liquid Legal Institute.

This panel selected the top three teams, who received cash prizes ranging from £2,500 to £10,000. Additionally, Problem Statement owners awarded prizes to the teams that best addressed their Problem Statement. The award ceremony marked the end of the Hackathon, leaving participants in anticipation of what's next.

The Winners

Following a close and challenging judging, the following winning teams emerged:

Precedent AI (First Prize) — enhances the efficiency of legal aid work by streamlining case review, research, and memo drafting. It can analyse extensive legal databases to provide pertinent precedents and citations, reducing a typical 13-hour workflow to less than 30 minutes. The social impact of Precedent AI is significant, especially in the field of immigration and asylum legal aid. It leads to improved access to justice, better quality of legal services, and a decrease in burnouts among lawyers.

Team members:

• Jacob Forward (King’s College, Cambridge)
• Xiaochen Zhu (Neo) (King’s College, Cambridge)
• Deepro Choudhury (Christ’s College, Cambridge)
• John Poole (Churchill College, Cambridge)

Nasir.AI (Second Prize) — helps businesses enforce the execution of their supply contract terms by using a two-stage pipeline to identify the contract terms, create executable code to monitor their fulfilment, and autonomously notify relevant stakeholders, such as payables teams and suppliers, continuously throughout the contract’s lifecycle. As a result, it provides an explainable, scalable, and private solution for businesses to measure the adherence to contracts, enhance contract monitoring, empower procurement teams, and aid in preventing supply chain disruption.

Team members:

• Zahra Farzanekhoo (King's College, Cambridge)
• Stuart Johnson (University of Cardiff)
• Vid Kocijan (Oxford, St Hugh's College)
• Hana Šerbec (Clare College, Cambridge)
• Robin Van Aeken (Hertford College, Oxford)

Discovery AI (Third Prize) — aids small and medium size businesses in resolving minor disputes like non-payment issues. Businesses can upload case data for analysis, and the AI will evaluate, organise, and estimate cost for legal proceedings. With a chat feature for inquiries, and connection with lawyers if legal proceedings are chosen, the solution aims to reduce financial losses from legal disputes and expand access to justice.

Team members:

• Zaver Dhanani (Trinity College, Cambridge)
• Krish Iyengar (Purdue University)
• Noella Spitz (University of Reading)
• Jean Zwerger (London Business School)

The Future?

LegalTech is here to stay! Following the Hackathon’s success, the King's E-lab plans to maintain and expand its LegalTech programming. Stay tuned by following our Hack the law LinkedIn page and/or subscribing to the E-Lab Newsletter.

List of participants

From United Kingdom (University of Cambridge, University of Oxford, Imperial College London, University College London, King’s College London, City University of London, London Business School, London School of Economics, University of Portsmouth, University of Surrey, University of Reading, University of East Anglia, University of Leeds, University of Manchester, University of St Andrews, Swansea University), Ireland (Trinity College Dublin), Netherlands (Maastricht University, TU Delft), Switzerland (ETH Zurich), France (University of Poitiers, SciencesPo Paris), Germany (Technical University of Munich, University of Augsburg, University of Göttingen), Austria (WU Vienna), Poland (Jagiellonian University), Slovenia (University of Ljubljana), United States (Yale University, Purdue University, UC Berkeley), India (Gujarat National Law University), and Australia (The Australian National University).

Previous
Previous

Law and Large Language Models, A Glimpse of the Future

Next
Next

Tackling the Hidden Costs of Food Insecurity