AI in Law – Evolution of Legal Research

Oh boy, have times changed!  Legal Tech has brought-in precision and accuracy to lawyers’ fingertips, which hitherto was available only to the top law firms and experienced legal practitioners. Law is still heavily dependent on human capital. However, with technology, the gap between experienced and young lawyers has reduced, at least in legal research.

Early-Internet Days

When I was in Law school, I used legal commentaries and volumes of books of case laws, to do legal research. Doing legal research was like a treasure hunt, we had to hunt from one clue to another across several books. Then, piece together a plausible story, interpret and apply law, to a legal issue.

I remember doing legal research for a Willem C Vis Commercial Arbitration Moot Court back in the 90’s. Access to the Internet was limited to now defunct internet cafés and browsing was at snails’ pace. To load even one page of #UNCITRAL rules and regulations took ages.

Mind you, Google had not yet arrived! In the pre-google days, we used engines such as Ask Jeeves, Yahoo! and Altavista to do our research. The experience was more cumbersome than doing research in a library surrounded by books. These search engines were good for generic research but not for legal research. We also had to contend with the fact that the content available on the Internet then, was limited in comparison to what is available today.

Beginning of Change

Then came a time when I started using a legal data base called Jurix. Jurix came on a CD ROM. We had to install the entire database onto our Personal Computer. The PC ran on Windows 95, and buggy 98. We had a CD Key in a Floppy Disk, which allowed Jurix to be installed only on one machine.

Jurix had 50 years of case laws, and came with four updates every year. Each quarter, Jurix would send us updates by post with newer case laws. The search was limited to a couple of phrases, and Jurix would throw up all the cases in the 50 years, featuring those phrases. We still had to browse through all the cases. And, we couldn’t just copy the case laws to Microsoft word, Jurix allowed us to copy the entire case law onto a clip board, and then we could type it in a blank document on MS word. That was one complicated tool.

Online Databases

Ten years ago, I was introduced to Westlaw, and LexisNexis as the Online Databases. Several lawyers could at once access the websites. And there was no end to information on case laws from across the world. This was truly revolutionary.

Westlaw and LexisNexis did not provide Indian cases, so I used them mainly to write research memos and briefs for law firms in the US and UK. Westlaw gave an option called Natural Language Search, and we could literally type a research query in English language. There was a very good chance of getting the case laws having all the search terms. In India we had online databases such as #Manupatra, #SCConline, etc., but they were not as evolved as Westlaw or the Lexis.

Getting There

Westlaw became #WestlawNext, the search became more intuitive. It suggested key terms to help lawyers look for answers. WestlawNext made legal research easy. The number of searches that one had to run; reduced. The tool looked for keywords from our searches, and at the same time pulled out cases that was very close to our query.

Here’s an example of searches in Westlaw and LexisNexis using natural language:

“spouse can restrain sale of community property bought from the common fund.”

The search engine would pull out cases using all the terms, but the first few results would have almost all the terms from case laws in a single sentence. This was a big step forward for online legal research.

The search engine at Westlaw was already good, the transition from Westlaw to WestlawNext was seamless, and felt like a natural progression.

Revolutionary

LexisNexis launched Lexis Analytics, an Artificial Intelligence (AI) driven legal research tool. LexisNexis acquired Machine Learning (ML) and Natural Language Processing (NLP) start-ups such as Lex Machina, Intelligize, and Ravel Law and integrated their capabilities into Lexis Analytics.

Lex Machina mines litigation data, giving insights about judges, lawyers, parties and subjects of the cases. The majority of the data provided by the program is optimized for patent information, copyrights, antitrust cases, and trademarks.

Intelligize offers a web-based research platform that provides content, news collections, regulatory insights, and analytical tools for compliance and transactional professionals. The platform can be used for securities and exchange commission filings, agreements and other exhibits, analysis and trends, regulatory materials, mergers and acquisitions, etc.

LexisNexis claims that Ravel Law can extract persuasive language from court opinions, challenges and motions. The language judge analyzes 100 motion types and examines millions of case-law documents to reveal powerfully persuasive language relevant to a case. Isn’t that amazing!

Westlaw has an AI based product of its own called the Westlaw Edge. The AI-powered version of KeyCite (citator) provides warnings when cases are no longer good-law which traditional citators would not have identified. WestSearch Plus, an AI-driven legal research tool, guides lawyers to answer specific legal questions, quickly. Integrated litigation analytics provides detailed docket analytics, covering judges, courts, attorneys and law firms, for both federal and state courts.

Here’s an example of the questions that one can ask on Lexis Analytics, and Westlaw Edge:

“How can I best challenge an expert’s testimony?”

There are videos on YouTube showcasing the use-cases of Lexis and Westlaw, and it is amazing.

Today, we have IBM’s Watson, which is essentially a cognitive (question-answering) computer system capable of answering questions posed in natural language. A couple of years ago, I heard some chatter that students from a university in Toronto, Canada, were developing and integrating legal databases with #IBMWatson, developing a capability to provide solutions to legal issues with case-law support. This is, essentially, “legal advice”!

It is now a reality! and it is called #Ross intelligence. With Ross, lawyers can ask research questions. Ross can read over a million pages of law, case laws etc, within a second, and give answers. Ross is currently limited to Bankruptcy law in the US, but the day is not far when all legal research can be done by a computer.

Today, most of the innovation in #LegalTech using AI is limited to the US legal system. Sometime soon in the future, with the kind of advancements we are making with homegrown technology, I hope the Indian legal system will also be able to use AI for legal research.