Very few know that the legal sector was one of the first to adopt Artificial Intelligence, with some corporate legal firms using it in some form since 2005. Even as the judiciary takes baby steps to digitise court work, lawyers across the country are using software like ‘Casemine’, ‘Mitra’, ‘Legitquest’, ‘Mike’ and ‘Kira’ for basic research, the sort of work which otherwise would have been handled by an entry-level legal associate. This analysis and retrieval of data otherwise takes an immense amount of time.
Huzefa Tavawalla, who heads International Commercial Law Practice at Nishith Desai Associates, allays the fears that jobs are at stake. “That can never happen because one of the things that a machine lacks is a conscience. This is one of the biggest challenges of AI,” says Tavawalla, whose firm was among the first to develop an in-house AI model.
Arup Roy, VP Research at Gartner, defines artificial intelligence as a programming technique which has its own language to break in business rules and knowledge into a software format so that a mathematical model, or algorithm, is derived to ingest data to output the desired result. After this initial stage, AI would “start learning on its own”. “That’s the level of intelligence which is still lacking. If the machine keeps learning over a period of year or so, it can probably go and identify the downfalls in matters and will make you aware how to go about it,” Tavawalla adds.
But developers and even users don’t think the Indian judiciary will be receptive to this new technology. Experts say that apart from a handful of judges, not many are tech savvy and might not be open to the idea of complete digitisation though it will help in speedy disposal of pending cases. Tavawalla says to start with all the paper filing can end. “In a pen drive I can have 10,000 pages. Intelligence is the next level. First the courtrooms need to adopt automation and then depending on how the acceptance is, move to the intelligence.”
In Cleveland and some other courts in the US, AI is being used for risk assessments of the guilty arraigned in court. Prior to pronouncing judgments, the courts are being guided by computer algorithms to decide whether a guilty should be permitted to return to his everyday life or not. Even in India, Aniruddha Yadav, founder of ‘Casemine’, says software can tell litigants that they are unlikely to win a case, and so offer other recourse. “This should really help the judiciary,” he says. At the moment, Tavawalla says, firms are using AI to scan contracts and pick those that could have issues. Around 100 contracts can be scanned thus in under three hours.
Searching precedents on specific legal points has always been one of the most time consuming job for all lawyers. Now, software tries to understand the facts of the case and list case laws specific to the needs of a lawyer. Anshul Gupta, co-founder of MikeLegal, says, “So where lawyers put in keywords to search certain case law, you can say our tool understands the law.”
While legal search engines like Manupatra and IndianKanoon are keyword-based content delivery platforms, AI understands the facts of the case on uploading a document and generates word suggestions to refine the search, lists laws mentioned in the document and provides visual aid of precedents, in addition to the case laws. Gaurav Srivastava, founder of ‘Mitra’, says that while it saves 40 per cent of the time, it is argumentative. “At the end of the day your ability to reach for right answers is your ability to ask the right queries.” At AZB & Partners, Senior Associate Abhimanyu Chopra says, AI is used “more like a smart check, instead of a spellcheck”.
For now, AI software will leave the inference of law in the hands of the lawyer, and focus on data crunching. “Lawyers can never be taken out of the equation. These systems only expedite the process of getting an information,” says Gupta.
While freshers could soon find it tough to get jobs just for data crunching, firms using the software say young lawyers should apply their intelligence in discovering legal points instead of delving into mindless paperwork. “In every revolution, people lose jobs but the greatest thing is they graduate to higher levels. Why should a lawyer spend time going from one library to another searching for the right book? He didn’t study law for that. The skills that a lawyer possessed in 10 years, they will start developing in the initial 4-5 years. It will happen in all industries,” says Srivastava.
Another factor is lack of regulations governing the developing technology. Sharmila Nair, a lawyer practising IP and IT matters in Kolkata, says AI is not limited to a certain geographical boundary and we cannot put a time cap on it, especially when we have started experimenting with the software. “In India we’re at a stage of reducing labour. This is the first part of AI. They are not looking at artificial general intelligence.”
Yadav too questions the future of unregulated AI, “For dispute redressal , the technology is here, but not the regulations.”
Nair believes that having an Act dedicated completely to AI would not be possible at this point in time, but a law on the ethics of using the software could be put into place. “One way to control this would be blockchain, it regulates who has used what to the last user. Linking blockchain to AI can help manage the accountability. If you are keeping AI or AGI as open source, this would be used as a timeline for who uses it next,” says Nair.
Addressing the practical issue circling the software, Chopra believes that as AI is at such an early stage, the cost of the software might supersede the need. “It is cost effective for firms with 300 to 600 employees, but an independent lawyer would not want to spend much money on such a software,” says Chopra.
Tavawalla explains that if and when AI works alongside judges, someone has to be held accountable in case of a faulty decision. “You can’t hold a machine accountable, can’t imprison it. It has loads of benefits and it’s going to be there for a long time but it might create some nuisance value.”