Law firms take baby steps in AI to increase efficiency and cut costs

In the first six months of her stint as head of artificial intelligence (AI) and Innovation at law firm Cyril Amarchand Mangaldas (CAM), Komal Gupta faced an uphill task. One part of her job was to assemble a team of lawyers who could work on the newly acquired AI tool, “Kira”. The other, and the tougher one, Gupta says, was to convince the partners at the firm to start using the tool and make them believe that it could be of help to them.

“Subscribing to a technology is the easiest thing a firm can do,” says Gupta. “The challenge lies in its safe implementation. We need specialised skills, interest and subject expertise to understand and train the tools. We also need to test the accuracy of the learning provided to the tool before rolling it out in the firm. This process has to continue throughout its lifecycle to achieve the desired results.” 

In 2017, CAM became the first law firm in India to embrace AI. Since then, the company has put together a team of six people who extract relevant information from documents so that it can be cross-checked against other standard clauses for compliance. 

CAM has been using its AI tool successfully. For example, lawyers working on a document review for a white-collar crime investigation used a tool to bring the number of documents to be checked down from 60,000 to 45,000 by avoiding duplication. 

“In another case, we needed to review almost 1,200 documents in one week while searching for specific criteria for a complex due diligence. Using the AI tool resulted in a 30 per cent rise in efficiency and a faster turnaround time,” the law firm said.

CAM is among a handful of Indian legal firms that are trying their hands at AI and machine learning. For most others, however, the technology remains an enigma. Though they recognise its benefits, many are still struggling with its application and use. And the situation is unlikely to change anytime soon, say AI experts at various law firms.

As Cyril Shroff, CAM’s Managing Partner, explains, “The legal practice requires strong problem-solving skills and emotional intelligence — skill sets that cannot be directly replaced by machines. In addition, we are operating in a complex and ever-evolving global environment. A single wave of tech solutions is unlikely to replace the legal landscape. After all, AI tools are based on human intelligence and are only as smart as we make them.” 

Mihir Parikh, strategic thought leader at Nishith Desai Associates, agrees. “AI is like a child. You have to train a child to talk in a particular way or not talk to someone. When children make mistakes, they learn. The same applies to AI,” Parikh says, adding that the lack of structured data in legal practice makes it tough to train AI in this field. 

Contextual information about cases is also important as it could help AI in coming to conclusions faster. This, too, is mostly unavailable. However, Parikh says that even if structured data were available, AI cannot be a quick fix for the solutions that law firms are seeking. 

At present, most law firms are using AI only for basic, routine work that are largely done by interns and junior lawyers. The more complex tasks which need analytical thinking and complex reasoning cannot be done by AI as yet, asserts Rajiv Luthra, Founder and Managing Partner at L&L Partners.

Hence, there’s no chance of a lawyer losing his or her job to AI in the near future. “It will take a lot more time for AI to entirely replace or even tangentially take up the host of activities that lawyers perform. Therefore, it would be prudent to view the use of AI in law firms as a tool to maximise efficiency and cut costs, rather than as a risk to lawyers,” Luthra says. AI could be equipped to perform the more mechanical tasks and act only as an assistant to decision-making lawyers — it cannot replace the decision-makers themselves, he adds.

Experts say that the technology is also unlikely to be adopted in the judiciary, at least not in the foreseeable future. Says Shivpriya Nanda, Partner at J Sagar Associates, “That does not appear to be a short-term reality. In the long term, there is definitely a possibility. How quickly AI is adopted by different courts will also depend on the scale and speed of digitisation."