Lawyers can breathe a sigh of relief: it seems even the most advanced robot is no match for a human when it comes to debating skills – at least for now.
Tomorrow, though, might tell a different story.
In February, the world’s top debater prevailed in a mano-a-nano cage match against IBM’s Project Debater – a computer described as the first artificial-intelligence system built to out-argue humans.
Harish Natarajan, grand finalist in 2016’s World Debating Championships, squared off against an AI system that IBM has been working on since 2012. “IBM Project Debater is designed to come up with coherent, convincing speeches of its own, while taking in the arguments of a human opponent and creating its own rebuttal,” according to USA New News. “It even formulates its own closing argument.”
To get there, IBM Project Debater was fed with more than 10 billion sentences of text from newspapers, magazines and other sources. It was also programmed to recognize and respond to nuances in a human opponent’s arguments.
The February contest – sponsored by the nonprofit Intelligence Squared US – was held before a live audience in San Francisco. The bot was not hooked to the Internet or given special programming. In fact, both contestants received the topic – should preschool be subsidized – only 15 minutes in advance. What followed was a more-or-less traditional debate, with each side presenting arguments, the other offering rebuttals, and each giving two-minute closings.
Audience feedback determined the winner. Though it was close, Natarajan won on points. AI’s most glaring weakness was its rebuttals. The machine tended to restate its own positions rather than specifically address opposing arguments.
AI and the Law
The debate made for entertaining headlines, but even the most ardent AI advocates caution against reading too much into the results. They say computers remain best-suited for repetitive, low-analytical tasks.
Legal experts concur.
“AI software will undoubtedly supplement some aspects of lawyering,” writes Nicole Black in Law Technology Today. “But most likely it will do so by allowing machines to do much of the tedious drudgery so common in some aspects of the practice of law, allowing lawyers to focus on higher level analytical work.”
Tracking and billing time is an area in which AI excels.
“AI can be built into a timekeeping solution to streamline and simplify the process of tracking billable hours, as exemplified by Intapp Time,” Black writes here. “Intapp Time captures time across all devices, including desktops, laptops, and mobile. The software then provides a daily summary which utilizes AI-type analytics to suggest relevant connections using other firm databases, such as names of contacts and cases based on the documents, etc. with which the user was interacting when the billable time was entered.”
How AI Can Help Out in the Law Office
In addition to time billing, nothing beats artificial intelligence for proofreading, catching and correcting errors, formatting documents, and organizing files.
Here are five other ways AI can come in handy in the law office.
- Predicting legal outcomes. An algorithm created by scientists at Chicago-Kent College of Law in 2014 successfully forecast the results of close to 8,000 U.S. Supreme Court cases from 1953 to 2013. Since then, the science has gotten even better.
- Contract review. Software products like LawGeex absorb new contracts that are uploaded into its database and match them with similar documents. Based on the comparisons, revisions and corrections can be recommended.
- Risk reduction. “TAR tools, including predictive coding, can be used to review information in real time,” writes Avaneesh Marwaha in Law Technology Today. “This allows lawyers to identify potential risks earlier, advise clients wisely about their exposure, and head off legal problems before they even occur.”
- Pattern recognition. Computers are great at detecting trends. In the law office, this can be useful in a myriad of ways, such as preventing cyber-breaches, tracking price fluctuations and protecting intellectual property. “For trade secret theft, we can identify behaviors that can quickly pinpoint the time frame the theft occurred, how it was accomplished and who was involved,” writes tech expert Jay Lieb in the ABA Journal.
- Legal research. An entire universe of law can be summoned with a single keystroke.
How does AI make your Law Life better – or worse?