Will artificial intelligence replace the lawyers?


When you think of a lawyer’s office, you typically visualize a cabin with row of books stacked up from top to bottom. With technology, however, lawyers today access entire libraries – judgments of courts of all the States and the Supreme Court and foreign judgments going back a hundred years– on the cloud. Now imagine if they didn’t have to read them either! ROSS Intelligence, an artificial intelligence based legal research software that gives answers to questions put to it in plain English. For instance, one could type in “is a non-compete clause legal in India? What is the position of the Bombay High Court?” and have Ross generate a response backed up with references including relevant judgments and readings. Ross will keep track of changes in law that may affect this response over time and keep you updated. Next you use Lex Machina, a legal analytics software, to mine litigation data from judgments to reveal insights about the court you are in, the judge hearing the matter, opposing parties and the opposing counsel – their win/loss rates, settlement rates, time to settlement, their experience with a particular judge, including the effectiveness of particular strategies and tactics. It can help uncover trends in damages, awards and remedies and inform strategy to file suits with the highest probability of success by comparing outcomes in prior, similar litigation, and even identifies which arguments prevailed and not, before a certain judge.

Take another example of Spotdraft or Beagle, that assists in contract review. Beagle reads your contract in seconds and highlights who the parties are, their rights and obligations, and liabilities. All this information is displayed visually with easy-to-read graphs and charts, so that key items of a complex document can be absorbed easily. Beagle learns your preferences and concerns over time and seeks to provide more relevant information with each use. Luminance, goes a step further and helps with due diligence exercises, commonly conducted for mergers and acquisitions or public offerings. It can read thousands of documents in a virtual data room and provide a summarized analysis of key issues, such as onerous liability clauses, one-sided termination provisions, extensive indemnities and other unfavourable clauses in a matter of minutes. ContractExpress can draft contracts, where the user has to answer certain dynamic questions about the nature of the contract and what the user expects of it, and the system can prepare an instant draft.

These technologies are, gradually, being put to use. As these technologies evolve and adoption increases, lawyers will be able to focus on more important cognitive tasks with better information and research at hand. Time and effort spent on routine tasks, mostly performed by junior lawyers, will reduce.

We believe that in no more than ten years, machines will perform routine jobs and certain mildly complex jobs (to a lesser extent, even client facing work) better than the average lawyer. Not every legal issue requires deep analysis and problem solving, or even if it can be done better that way, the man hours and cost such research demands make it commercially unviable. As long as AI can provide reasonable quality with consistency at high speed and low costs, it would already be performing better than human lawyers of today in solving the average low-stake legal problem. Human lawyers will continue to add value with their creativity, artfulness and empathy, lacking in today’s AI. They will need to move up the value chain and grapple with more complex, newer problems. We cannot yet imagine AI doing complex drafting, even simple negotiations and strategy formulation. Of course, theorists, and even technologists are already talking about achieving ‘singularity’ in the longer term – the ability of intelligent software to self-learn and self-improve in a small period of time, achieving infinite intelligence. For the purposes of this blog, we will not go that far, however are tempted to remind our readers that IBM’s Deep Blue defeated Gary Kasparov at Chess twenty years ago. Agreed that Chess is a rule-based game, and the human environment is anything but that, but these systems will rapidly improve and in the meanwhile will be able perform very advanced tasks in any environment where it is possible to lay down the rules of the game.

Is it dooms day for lawyers and law students? AI will also present opportunities. A lot of jobs will be created for lawyers who can also code or can work with coders. These jobs will be with technology companies. Close to 350 legal-tech start-ups have raised about Rs. 6000 crores in the last five years globally. This is a trend one cannot ignore.

For a discussion on trends in legal-technology and artificial intelligence, please feel free to reach out to the Ikigai Law team.

the status quo

Bringing what's next...