Affordable legal advice for all – from a robot

An academic and a lawyer have teamed up to develop a robot lawyer, which, if successful, will make legal advice affordable to people from all backgrounds, while revolutionising the legal sector.

Robots could take on significant parts of a lawyer’s work, reducing the costs and barriers to access to legal services for everyone, rather than just those who can afford the high costs.

The project, at the University of Bradford, is initially working on a machine learning-based application to provide immigration-related legal advice, but if successful, it could be replicated across the legal sector.

The idea has received government backing in the form of a £170,000 grant from Innovate UK Knowledge Transfer Partnerships. Legal firm AY&J Solicitors is providing a further £70,000 as well as the vital knowledge of lawyers.

The project was devised by Yash Dubal, immigration lawyer and director at AY&J, and Dhaval Thakker, associate professor at the faculty of engineering and informatics at the University of Bradford. It will harness complex knowledge graph technology and deep learning algorithms to analyse case law and learn from it.

Thakker specialises in explainable AI, which is programmed to describe its purpose, rationale and decision-making. He said the idea emerged when he and Dubal had a conversation about where AI could be used in the legal industry.

“There are a lot of places where AI is already being used in the legal context,” he said. “Most is around contract review and analysis and there are many companies and systems out there that do this, but it is heavily US biased.”

The project aims to create a robot assistant that will be able to do 20-30% of an immigration lawyer’s job.

Thakker said much of the early work done by lawyers in immigration cases could be taken over by AI. “When immigration applicants come to the company and need information before they sign up, they want to know which immigration category to sign up to and whether they have a case for this,” he said.

“A lot of time is spent helping to answer those questions before they sign up as a client. Then, when they do sign up, there is a lot of paperwork in terms of getting all the information required.”

In the project, users will interface with a conversational agent system called LILA (Legal Immigration Artificial Intelligence Advice), which uses explainable AI so that the decisions it makes are presented in a way that humans can understand.

LILA will capture and analyse knowledge from solicitors, experts, statutes and case law. It will constantly monitor decisions made by existing experts and learn from them. It will collect information, ask questions of clients in the form of text or voice and will even be able to prepare a technical legal response.

Thakker said solicitors use their knowledge and experience from previous cases to help clients, but because case information is documented, a computer could use machine learning to do this.

“We are looking at a search engine that uses machine learning and deep learning,” he said. “The project is using knowledge graph technology, a visual way to represent the thought process of intelligent technologies, to ensure the system does not become a ‘black box’.”

Thakker, who has researched explainable AI and knowledge graphs for 15 years, added: “There are a few ideas, but it is early days. One of the ideas is to build a search engine to help solicitors find cases. You have a candidate in front of you and the system can use that profile to find similar cases.”

The initial AI consultation would be offered for free, which supports the project’s main aim. Eradicating the high costs of legal advice and making it available to people in all social groups was the driving force for AY&J’s Dubal, who specialises in visa and immigration services to businesses and individuals.

“If you look at the legal profession, lawyers cost £200 to £300 an hour,” he said. “Even in a rich country like the UK, ordinary people can’t afford that.”

Dubal is not concerned that legal firms will lose revenue if a significant part of their service is automated. “Although I am a lawyer, my focus is the problem of ordinary people, not the legal profession,” he said.

“I come from a very humble background and my genuine intention is to demonetise the entire legal industry and make proper legal advice available to everyone. We want to provide legal advice to everybody, not just the few who can afford it.”

Dubal is not your typical director of a legal firm. After completing a computer engineering degree in India in 2002, he left for the UK “for a better life” alone, with no contacts and no support.

He worked as a cashier at Smithfield Market and found a place to live in east London. The first £50 he saved, he send to his mother so she could buy her first fridge.

“I worked hard, at times 18 hours a day, seven days a week, with no holidays,” he said. “I worked as a shopkeeper and security guard and ran a small college in London. I started helping friends on their college admissions, from which I earned small commissions.”

Dubal chose to submit his immigration application to the UK without legal advice – and failed. “I lost hard-earned money and equally precious time,” he said. “I had to hire a lawyer at significant cost in order to get a successful outcome.”

He then decided to help others with their UK immigration applications. In September 2008, he set up an immigration law firm from a corner of his bedroom in a rented property in Hounslow.

“Our reputation spread, and by 2013 we moved to offices in Chancery Lane, central London,” he said. “Today the company, AY&J Solicitors, is a Legal 500-ranked company.”

The reason the AI project is starting with immigration applications is because of Dubal’s expertise in this area. “My job is to provide knowledge to help create the application through machine learning,” he told Computer Weekly.

The team does not expect to replace humans in the near future, but is trying to create a semi-lawyer assistant to reduce the need for human lawyers. “The idea is to create a semi-lawyer,” said Dubal. “This is a humble effort and if we take a step forward, one day we will reach that.

“It could demonetise the entire sector and in so doing, it will reduce the need for specialised skilled lawyers, and that will reduce costs. This could benefit tens of thousands of people.”