Will AI replace lawyers and judges or simply change the way they work and think about the law?
Back in the late 1980s, lawyers used to type their legal briefs on old-fashioned typewriters. Briefs were shorter then but drove right to the central issue of the case. Now some lawyers copy/paste long passages from other briefs and compose new ones running to hundreds of pages. Judges’ opinions are longer too, burying the true reasons behind their decisions in page after page of verbiage. Word-processing software makes it easy to spin out long briefs or opinions where the real issues often become blurred. More importantly, it seems the way we write has also affected the way we think.
This has happened because of the technological shift from writing in longhand or typing, to computers, printers and amazing legal databases like LexisNexis providing finger-click access to all of the country’s law. This revolution has changed the way all of us work. It has certainly eased our work. It has also probably widened access to legal services. This technological leap, behind us now, also has its downsides. To name just one culprit: voluminous submissions to courts struggling to identify core issues in the case.
The legal world has taken action to defend itself against the downsides of these new technological developments. Courts have imposed page limits on briefs and issued practice directions on how they should be construed. Bars have begun rigidly enforcing ethical standards for proper provision of legal services, with due respect for cost constraints and proper allocation of judicial resources. Bars and law schools are teaching how to write effective court documents, and these skills are tested on the bar exams. These measures are now standard in countries like the UK or the US and to some extent are also being introduced in other countries such as Poland. The legal community has become acutely aware that for limited judicial resources to be used reasonably, judges should only focus on the truly contentious issues. Technology helps lawyers do their jobs, but it must be used in a proper manner and for proper purposes.
We are now on the verge of yet another technological revolution: the use of artificial intelligence to assist lawyers in their daily routines and maybe even replace them in areas were tasks are repetitive and relatively risk-free. We believe that this change, like the previous ones, will ultimately be good for lawyers and the justice system overall. But it must be implemented cautiously to avoid disrupting the whole justice system.
In many jurisdictions computer-assisted document review has been approved as a sufficient substitute for review by a human lawyer, if done properly. However, the latest technological advancements go much further than this. One US law firm recently hired a new “associate”: ROSS, an IT legal assistance system based on IBM’s Watson AI. Simply put, ROSS aims to replace lawyers doing legal research. ROSS and other technologies like it intend to rely on deep learning, reinforced learning, and, in the future, quantum computer technologies. This all means that computers will be able to solve problems that have not been previously programmed into their systems and thus foreseen by their engineers. ROSS and similar systems teach themselves how to solve problems—in the context of this article, specifically legal problems. One of ROSS’s developers, Andrew Arruda, sees this advancement as an opportunity to cut legal bills and thus provide wider access to legal services and justice than before, in particular for those who cannot afford a lawyer. Arruda is right. ROSS and other technologies will revolutionise not only the way lawyers work, but also the legal market. So far, AI systems are not sophisticated enough to actually replace lawyers. They only free them from the most repetitive tasks. But the ultimate aim of these AI technologies is more ambitious.
The developers of AI systems in the legal sphere aim to let computers serve justice. They claim that this will be cost- and time-efficient, resistant to corruption, and ensure more equal treatment by eliminating the bias inherent in human judges. To a certain extent that does seem possible. It is obvious that the latest AI technology can already replace human lawyers in certain simple types of legal work. It seems that certain matters could just as easily and justly be resolved by computer systems analysing specific sets of data and following certain rules. For example, advice on inheritance or company registration issues could be performed entirely safely by a computer. All that is needed is to input the right data to be scrutinised by the system before disclosing the results of its analysis. If these new systems are reinforced by other breakthrough technologies, such as distributed ledgers (e.g. blockchain), this could generate further efficiencies. If public records such as land registers were run on blockchain, the new computer system could easily, automatically and in paperless mode update the title to a decedent’s property. The same could apply to complex transactions based on smart contracts or smart wills, which would not only generate binding legal effects but also self-execute using new blockchain technologies.
While the latest AI systems can substantially assist lawyers in their work now, we are confident that in the near future they will be ready to replace them. Computers would not only do this work more quickly and cheaply, but often probably better. Just imagine how many resources could be reallocated from the document review accompanying every transaction to more skill-oriented tasks. Computers can review all the documents, assess the risks, and produce reports with final recommendations just as well as human lawyers and considerably faster. As transaction work generates large profits for law firms and constitutes a major cost item for businesses, it is not difficult to imagine how in the future the market will simply force law firms to replace lawyers or paralegal document reviewers with more cost-effective machines. And hundreds of other tasks soon could also be replaced by machines: client interviews, document collection, etc.
Ultimately, we believe, computers will be capable of learning what the case is about and what information is necessary to solve it. The computer would interview the client, collect all documents and electronic files, draft the briefs and make the filings. And probably another computer at the court would analyse the materials and deliver a judgment. It seems to us it is a question of “when” this will happen, not “if.” We will experience such computer-driven justice systems shortly, and these systems will inevitably replace some lawyers. But they will not eliminate them from the system entirely—not now and not in the future. This is because all these machines lack a feature that modern justice systems are intrinsically built upon: empathy. Since Roman times, our societies have tempered the stern maxim dura lex, sed lex with a more equitable approach to legal issues. Civil, criminal, family and constitutional law require human empathy for truly just application. Although much of a lawyer’s work can be successfully done by machines, for this reason at least we hope that human lawyers will in the end always be needed.
While human lawyers are intrinsically necessary to justice systems, AI technologies will all the same revolutionise the way we work. Law firms will no longer need hundreds of junior associates for legal research. That will cut operational costs and ultimately decrease the total fees paid by clients. The downside of this is that hundreds of law students will struggle to find work. Law firms in turn will need to reinvent how they train their young lawyers. Traditionally, junior associates start from conducting legal research, reviewing documents and doing repetitive work, with a view to learning the practical aspects of providing legal services. These tasks will more often be done by machines, so junior lawyers will have fewer opportunities to train and develop their own skills in this respect.
We believe that law firms need to prepare for all these changes. It would be unwise for them to ignore the latest technological developments. The technological shift is unavoidable. Firms must hire IT specialists to implement new systems in lawyers’ daily routines. At the same time, firms must assist their lawyers in learning technology and understanding how it works. Firms and bars must offer training for lawyers in using and “cooperating” with “machine lawyers,” teaching them how to freely and simply talk about technology. Law schools, bars and law firms should also understand that the most valuable lawyers will be those who are well-rooted in human rights, legal philosophy and human psychology.
The future of law is for lawyers who are able to demonstrate the human aspect of all cases and the core principles behind them. All the rest will be done by machines.
Łukasz Lasek, Jakub Barański