AI Enters the Legal Profession: The First Step Is to Become a Lawyer’s Right-hand Man
For a long time, the impression of lawyers on the public is that they are elite, eloquent, and highly paid, but the hardships behind this job are rarely seen. In addition to appearing in court, lawyers have to spend a lot of time and experience on legal data research, contract document review, etc. For this reason, technology companies have launched AI lawyers to share these repetitive and mechanical tasks. The legal industry is also facing a transformation brought about by AI.
Some time ago, the first domestic legal workplace observation reality show "The Heart-Stirring Offer" revealed a corner of the legal industry to the public, attracting a lot of attention and discussion about this industry.
Through this program, the public has the opportunity to see lawyers’ daily work status up close, and also understand that behind the stereotyped impressions of “elite” and “high salary”, there are the complicated work and heavy pressures that lawyers face in their daily lives.
Behind the high salary, there is also complicated basic work
An industry insider said, "The profession of a lawyer is all about exchanging time for money." The 996 pace is not only the norm for programmers, but also for lawyers.
The most common thing we see lawyers doing is negotiating in court and fighting for clients. But in fact,Lawyers are divided into litigation lawyers and non-litigation lawyers, and their work content is also different.
Litigation lawyers accept clients' commissions to help them fight lawsuits. In addition to defending in court,There is a lot of preliminary work to do: reading files, writing petitions, collecting evidence, and studying legal materials. Some major cases may have dozens or even hundreds of files.
Non-litigation lawyers basically do not appear in court. They are responsible for verifying various materials and making various document modifications. The results of their work are various documents, legal opinions, and agreements.
The profession of lawyer has a long history, but to date, its working model has been relatively lacking in innovation, and some management software and tools are still relatively rudimentary.Therefore, lawyers actually spend a large part of their time working at their desks, dealing with massive amounts of documents, information, and contracts.The rigor of the law also requires lawyers not to be negligent in the slightest, which invisibly adds to the pressure.
An intellectual property lawyer said that in fact, most documents, whether civil agreements, commercial contracts, or dispute prosecution documents, are similar. Therefore, can this part of repetitive and mechanical work be handed over to AI? The AI lawyers that have emerged have already demonstrated their strength in this regard.
Contract review is handed over to AI, which is highly efficient
On December 4, a special legal skills competition was held in the small auditorium of Zhejiang University Law School.
because,In addition to 16 human lawyers and law school students from well-known law firms, the competition also included an AI.
Competition information at a glance:
Competition topic:Contract Review
Contestants:Human group, AI group, human-machine collaboration group, 5 teams in total
Competition requirements:The participating lawyers reviewed 5 rental contracts totaling nearly 20,000 words within 30 minutes and found the risk points in the contracts (a total of 50 risk points were preset in the contracts)
Results:The human-machine collaboration team won the championship
After the game starts,The AI lawyer completed the review calculation in less than 1 minute and automatically submitted the paper.This speed can be said to be a true read-through, catching everyone off guard. Not only is the speed fast, the AI lawyer also has a high accuracy rate, reaching 96%.
However, the final championship was won by the human-machine collaboration team. After the AI pre-screened the contract and reviewed it with human lawyers, it found more difficult and risky points.
The "AI Lawyer" participating in this competition uses Alibaba Damo Academy's deep language model system ALICE, which enables machines to understand human language that contains complex logical thinking.
at present,In addition to contract review, this AI lawyer also has nearly ten legal skills, including court record keeping, assistant trial, and document quality inspection.
PPI claims and agreement review, AI has already mastered
In fact, AI technology has been applied in the legal industry for some time. It has defeated humans in competitions many times and has therefore been recognized by major law firms.
2017: PPI Claims Prediction Competition
As early as November 2017, Case Cruncher Alpha, an AI lawyer created by four law students from the University of Cambridge, participated in the first AI vs human lawyer competition in history with 100 lawyers from several well-known law firms in London.
Competition information at a glance:
Competition topic:PPI Irregular Sales Claims Forecast
Contestants:AI Lawyer Case Cruncher Alpha vs. 100 Lawyers
Competition requirements:Judging whether the Financial Ombudsman will allow claims based on hundreds of examples of PPI mis-selling
Results:AI beats humans with 88.6% accuracy (66.3%)
In this competition, participants submitted a total of 775 predictions for these cases.Case Cruncher Alpha outperformed human lawyers (accuracy 66.3%) with an accuracy of 86.6%.
2018: NDA Review Competition
In February 2018, LawGeex, an Israeli technology startup, teamed up with Duke University, Stanford University, and the University of Southern California to hold a human-machine lawyer competition.
Competition information at a glance:
Competition topic:Confidentiality Agreement Review
Contestants:AI Lawyer LawGeex vs. 20 Top American Lawyers
Competition requirements:The contestant is required to review five (14 pages in total) non-disclosure agreements (NDAs) within four hours and point out 30 legal issues, including arbitration, relationship confidentiality, and compensation. The scoring criteria are based on the accuracy of the contestant's corrections.
Results:AI 88.6% accuracy beats human 66.3%
The end result is,AI lawyers defeated human lawyers with an average accuracy of 85% with an average accuracy of 94%.In one of the confidentiality agreement reviews, the AI’s accuracy rate reached 100%.
In terms of time consumption, AI is also far less than humans. It can complete a task in 26 seconds, while the fastest human needs 51 minutes.
Therefore, whether in terms of accuracy or efficiency, AI has a clear advantage over humans. Every piece of data in the competition results shows that AI is capable of doing this job.
In law firms and courts, AI becomes a lawyer’s right-hand man
Almost since their advent, AI lawyers have found jobs in law firms.
In 2016, ROSS, known as the world's first AI lawyer,He was hired by Baker & Hostetler, a well-known American law firm, to assist in handling the company's bankruptcy-related matters.Since then, ROSS has won the favor of more than a dozen large American law firms and has been used to answer specific cases and legal issues.
The LawGeex AI lawyer, which once won the man-machine competition for confidentiality agreement review, has also been used by well-known law firms and accounting firms.
In China, the Shangcheng District People's Court of Hangzhou introduced AI judicial assistants in 2019 to assist in presiding over trials.With the assistance of AI, everything from court transcripts to the generation of judgment documents can be completed automatically, greatly shortening the case closing time.
It has to be said that AI lawyers have already brought about a revolution in the legal industry. However, it is still too early to say that AI will truly replace lawyers.
AI is responsible for basic work, while humans are responsible for complex business
From the current technology point of view,AI only replaces repetitive and mechanical work, and it will still take time for it to understand the semantics of the files.
Therefore, AI lawyers are currently still mainly engaged in basic legal research, helping humans save time on reviewing and auditing information, so that they can spend their energy on things that require human creativity, judgment, and execution.
According to data from ROSS Intelligence, the research and development organization of AI lawyer ROSS,In a lawyer's work, one-fifth of his time is spent on basic legal research; each year, law firms across the United States invest approximately $9.6 billion in legal research.
The introduction of AI lawyers to perform basic legal research not only saves the company huge costs, but also speeds up the judicial process and greatly improves case handling efficiency.
The advantage of AI lawyers is not only to save time and manpower costs for law firms, but alsoFor C-end users, the price of consulting with AI lawyers is much cheaper than that with human lawyers.
Take AI lawyer ROSS as an example. It is aimed at C-end customers and can provide problem retrieval, case handling, and case summary services for only US$79 per month. Compared with the average consulting fee of US$200 to US$300 per hour for American lawyers and RMB800 to RMB6,000 per hour in China, AI lawyers are much more cost-effective.
For all these reasons, AI lawyers will become a powerful tool for practitioners in the legal industry, and in future law firms, the human-machine working model may become the norm.