Artificial intelligence and the risk of uberisation are frightening, rightly and wrongly
Artificial intelligence makes dream, fantasize or even nightmare men of dresses. When she is interested in the exercise of justice and transforms their profession, she can also make them jump. Legaltechs – these disruptive young start-ups thanks to high-tech – continue their powerful growth, presenting each month projects aimed at improving, if not revolutionizing,the way of doing justice.
Knowing that it is being monitored, AI presents itself today as an ally in the service of professionals,
Artificial intelligence makes dream,fantasize or even nightmare men of dresses. When she is interested in the exercise of justice and transforms their profession, she can also make them jump. Legaltechs – these disruptive young start-ups thanks to high-tech – continue their powerful growth, presenting each month projects aimed at improving, if not revolutionizing,the way of doing justice. Knowing that it is being monitored, the AI presents itself today as an ally in the service of professionals, especially lawyers. Among its promises and benefits: streamlining legal research among a wealth of case law, freeing up firms for tens of hours to devote themselves to their true added value, client advice and strategy.. But already, other forms of AI promise in particular to statistically predict judgments given those made in the past. Predictive justice provokes the ire of a judicial world unprepared for such a rapid explosion. 4.3 million court decisions are made each year in France. A volume that makes law a living material, nuanced with a thousand special cases, always embellished with new contexts. A mass of documents, often long, technical, repetitive: in short, an ideal playground for a “datamining” tool.
This is the reflection that some start-ups overdosed on new technologies, stimulated since 2016 by the law Lemaire for a Digital Republic that makes in theory judicial decisions accessible to all. Their finding? Lawyers spend (understand waste) many hours searching for and analyzing texts that a well-trained machine could sort much faster, without omitting any information. Automation saves time. “For a lawyer, it’s impossible to read everything anyway,” says Louis Larret-Chahine, co-founder of Predictice, a start-up founded in 2015, which promises to facilitate legal research and analysis, “in , “a way comparable to that of a medical scanner, which no one felt was questioning the doctor’s expertise.” l’expertise
This optimized search engine, capable ofanalyzing 2 million documents per second, is already implemented with more than 2,000 French lawyers, including many of the major firms.
What AI are we-t- talking about?
While the emergence of legaltechs is rapid – especially in France, where fundraising for these start-ups has jumped by 92.2% in one year – the day is quite distant when a robot will deliver its first verdict. “Today, AI is 90% fantasies and 10% real,” says Bernard Lamon, founder of New World Lawyers, a law firm specializing in new technology law and co-author in 2017 of the report ‘Future of the Lawyer Profession’. For him,the new tools are welcome, allowing to work “sometimes faster, sometimes better”,and finally to have intelligent engines to fluidize the research work in a natural language, or almost.
In fact, the vast majority of start-ups in the sector do not really rely on artificial intelligence, or else in its weak sense, focusing on the repetition of a known behavior, in the restricted field. Document en analysis, editing of contract Analyse templates, carrying out formalities online, connecting with legal professionals, first level of legal advice, sending letters, amicable raises… relances A trend on which BtoC-oriented companies are surfing, which are increasingly known to the general public: Thisisalf.com, Legalstart.fr, Litige.fr, Demandezjustice.com… Solutions certainly legal, but whose technical dimension remains limited. At the higher level, AI appears mainly in the form of automatic natural language processing. There are or Doctrine platforms, dedicated to legal professionals, capable of sorting for them in the growing mass of court decisions. (The whole thing for a price ranging between 130 and 190 euros permonth.) mensuels They relieve them of thankless tasks and allow them to focus on what makes their legitimacy and value: advice, human accompaniment, understanding the context, defining a winning strategy.
Predictive justice: the scary horizon
Only, the AI can much more, even if it makes jump. With a point of tension that has become a symbol of this debate: predictive justice. “The advent of these commercial enterprises proposing to predict justice has shocked many people,
The day is apparently quite distant when a robot will deliver its first verdict. “Today, AI is 90% fantasies and 10% real””
The vast majority of start-ups in the sector do not really rely on artificial intelligence, or else in its weak sense, focusing on the repetition of a known behavior, in the narrow field. Document en analysis, editing of contract Analyse templates, carrying out formalities online, connecting with legal professionals, first level of legal advice, sending letters, amicable raises… relances
confirms Mr. Olivier Fontibus, Chairman of the NBC Law Exercise Committee (National Bar Council). barreaux). In a way, this debate has been beneficial,because it has allowed us to trigger an in-depth reflection on the impact of technology and then to react.”
Predictive justice – a mistranslation of its American version, “predictible justice” or , “”predictable” justice- is the stated ambition to provide,on the basis of thousands of past judgments, the verdict that such a jugements court could deliver, at any given time, on a particular case.
“Aren’t the first questions a client asks his lawyer: what are my chances of winning,and how many damages can I expect?” says Louis Larret-Chahine.-Chahine. Predictice thus proposes to provide law firms with a riskreport, summarising the possible gains or losses in relation to past decisions. .
The reduction of a judicial verdict to a calculation of probabilities, by definition devoid of any human or contextual element, has provoked such an outcry that the start-ups concerned have been urged to be cautious. Predictice, whose name hardly hides this vocation, or Case Law Analytics, which clearly shows its ambition to “model the judicial decision process to present all the decisions that would be made on a given file”, are the first targets.
“Predictive justice has the principle of passing a large amount of data to the reel of an algorithm. But who controls -this one? No algorithm is neutral! ” en warns Olivier Fontibus. As for the implementation of such tools to come and “increase” the magistrates by providing them with their analyses, it would, according to him, pose a threat to the very idea of the verdict: judgment on criteria that are not transparent or even obscure,an exaggerated conformity to what has already been judged… For the time being, however,the technology has not appeared sufficiently developed to call into question the functioning of the institution. “We got scared but we are pretty reassured,” says Olivier Fontibus. The latest experiments show that predictive justice is stronger in communication than in anticipation.”
Open data: battles in a mess
Predicting the verdicts, Grégoire Tomas, VP Marketing and Growth of Doctrine, is careful: “Predictive justice is a marketing concept that has had its effect.. In reality, it does not work, and above all is not desirable for anyone, “tranchet Start-up created only 3 years ago, Doctrine already gathers 100 employees, and shows a growth of 10% per month with 5 new customers per day. 80% are lawyers. Engine of this success? The collection of the most valuable resource: data. “Our first job is to centralize all legal information, now public. Our addedvalue? value this information, make it relevant,and assure our customers that no data related to their business escapes them,” explains Grégoire Thomas.
Although it does play by the oracles, Doctrine is the object of questioning,and even a complaint filed by the NBC and the Paris Bar. At issue is the means by which the company would collect these famous data. A conflict that has its roots in 2018, with a resounding case . Accused of having “typosquated”(using a domain name strongly resembling that of an existing site, to create confusion) the names of famous law firms and universities,Doctrine would thus have collected documents that it could not have obtained otherwise. At the time, the company denounced the mistake of an intern,” recalls Olivier Fontibus. The NBC now believes that some of its methods need to be studied more closely by the courts. Meanwhile,its competitor Predictice has accused Doctrine of having typosquatedit, en using a domain name too close to its own…
These battles can be explained by the blurring of open data in court decisions. Since 2016 and the enactment of the Lemaire Act, the implementing decrees relating to this particular point have still not been promulgated (they should be enacted this year). The main brake? the difficult anonymization of decisions,namely the concealment of names or elements to identify the parties, judges or even personal of the registries, in order to protect their privacy.. While waiting for the decree,the breach continues to be exploited, and the data guerrillas to continue.
Threat to the employment of lawyers?
Another uncertainty accompanies, wherever it takes place, the arrival of AI. Automation, and therefore potential job cuts for lawyers. Bernard Lamon laquelle refuses to give up a nos fear: Lamon : “We are as rarely a part of our working methods. We’re all going -t-to have to move,”he said, recalling the population growth of a profession that has grown from 30,000 to 70,000 people in 20 years,and which, according to him, still has great prospects ahead of it.
Nevertheless, legaltechs have already invested a part of the legal market, through platforms like Captain Contract, allowing to obtain a formal contract for a few tens of euros instead of several hundred, or even thousands.. Others, such as DemandezJustice.com, offer help to win cases in small disputes for which consulting a lawyer would not be profitable… Classical disruption is at work,and the Uber syndrome lurks: “Our situation has some similarities with the uberisation of taxi drivers, concedes the chairman of the NBC’s Law Exercise Committee. For a long time,we have not occupied the field of small files, we must admit.. It is in this area that many new players have positioned themselves.” Logically, to be delighted with a market on which we did not work, or so little, should not upset the dynamics of employment. Unless other elements disturb it… A specialist in new technologies, Bernard Lamon is closely following legaltech projects, which he predicts” new wave of innovative services is about to enter the market”.” Probably bringing as many new legal puzzles and societal issues, for sure deeper than a game game of go.
“For a long time,we have not occupied the field of small files, we must admit.. It is in this area that many new players have positioned themselves.”