Here Comes the AI Lawyer

AI-driven legal assistance software is gradually becoming a routine tool among lawyers. How effective are they?

After making its mark on several industries, Artificial Intelligence (AI) is now all set to enter the legal arena in a big way. AI-driven tools are gradually taking over legal tasks that raises hopes of clearing the huge backlog of litigations that lie pending at courts all over the world – owing to an acute shortage of legal professionals. Of course, it is too early to hope for a robot lawyer accepting your brief anytime soon, but the developments are promising. And more and more firms – big and small – are jumping into the bandwagon.

The origin

The humble DoNotPay app is being addressed as “the world’s first robot lawyer” by its developer Joshua Browder. Living as a teenager at London, Browder was plagued with regular parking tickets. At the age of 18, he developed the app to help him contest the tickets. From his own experience, Browder discovered that if he could logically contest these tickets in court using the right legal vocabulary, some of these tickets might be withdrawn if his point is justified.

However, that would need the services of a lawyer. you can save a lot of time and money. This was when he developed the first version of DoNotPay just within a few weeks in 2015. Instead of copy-pasting the same appeal every time, his app suggested the best legal language to use depending on the case at hand. It worked for Browder, and now at the age of 24, he has set up his own company at Silicon Valley in California to commercially develop it.

DoNotPay now claims to have 150,000 paying subscribers. They can simply type in their side of the argument in the application chatbot, and using a machine learning model the software suggests the legally appropriate terminology to express the argument. It supports composing appeals for a wide range of problems: trade-related complaints, visa applications, insurance and indemnity, refund for cancelled holiday booking or membership – being the most common of them.

It’s catching on

Luminance is another firm that has developed similar AI-based legal assistance software. It can be used to quickly browse and sort massive quantities of case documents. As reported by BBC, London-based barrister Sally Hobson, representing the law firm ‘The 36 Group’, had recently used the Luminance AI tool for a complex murder trial that required speedy analysis of over 10,000 documents. The tool did it four weeks ahead than it would have taken to complete with manual effort, also saving £50,000 in the process. More importantly, this software is being used by over 300 other law firms in 55 countries, working in 80 languages.

The legal software Litigate, developed by an Israeli firm, can not only help in sorting documentary evidence, but can also search for any relevant legal precedents to help lawyers prepare and structure their cases. London law firm Taylor Wessing is now using this software on a regular basis. Speaking to BBC reporters, Laurence Lieberman, who heads, Taylor Wessing’s digitising disputes programme, explained how convenient the tool is. According to Lieberman, once the case summary and the related pleadings are uploaded, the tool will plot out the key players, link them together, and draw up a chronology of the key events and explanation of what happens on what dates. That sounds impressive indeed!

Deloitte Legal, the legal arm of Deloitte – one of the Big-4 consultancy firms – is also ready with its TAX-I software, which can analyse historical court data to search for similar tax appeal cases. And then it can predict the outcome of current appeals based on historical verdicts. The firm claims an accuracy rate of 70%. But what help will lawyers get from this? According to Deloitte, a reliable and quantifiable prediction of the likelihood of success can help lawyers to determine whether the case should be pursued.

So what’s the verdict?

Overall, AI-driven assistance software is gradually becoming a routine tool among lawyers and gaining recognition. In 2020, DoNotPay won an award from the American Bar Association.
Although critics argue that software-generated legal advice is not yet accurate enough to act upon, improvements are happening every day. DoNotPay claims an 80% overall success rate. Deloitte says their TAX-I is 70% accurate. Eleanor Weaver, chief executive of Luminance, is of the opinion that although earlier-generation document-checking technologies were no better than keyword searches, current sophisticated software are adept in connecting associated words and phrases.

And these can be of immense practical benefit in countries that suffer from a shortage of legal professionals. Brazil had a backlog of over 100 million court cases. How long could that take for human judges and lawyers to dispose of, if ever? It might be more prudent looking for a reasonably refined automated legal system instead to predict the verdict through analysis and clear the humongous backlog.

Acknowledgement:https://www.bbc.com/news/business-58158820