Justice in the age of artificial intelligence and big data: 🤔Think about the future

How far will robots and artificial intelligence go in the legal profession? How the justice in the age of Artificial Intelligence and big data will effect?

If automation and technological acceleration are changing the game, certain predictions are still science fiction for experts. “Predictive justice,” software-assisted contract execution, legal firms recruiting robots etc.  Many dramatic statements in the last few years have suggested that the legal and judicial professions will soon be uberized due to technical and digital developments. When it comes to legal innovation, there are a few that have the potential to revolutionize the practice of law while others have an exploratory or even fantasy-like quality.

There are limits to these systems when matters are complex or confusing, else the profession of legal counsel would have been replaced by a simple search on the Internet.

Some aspects of “predictive justice” are similar to those of traditional justice. It has been increasingly common in recent years for various algorithms to be devised to assess the likelihood of a case’s victory and to estimate the damages that may be awarded. It is not a question of ‘replacing’ lawyers, contrary to popular belief, says Thomas Saint-Aubin, a Sorbonne Institute of Legal Research associate researcher and Seraphin Legal director of research and development.

The goal is to make it easier for people and businesses to conduct their own research, as well as to relieve specialists of tedious work so they can “concentrate on the most interesting part i.e. the analysis of the file, the creation of an argument, personal contact etc. Thom Saint-Aubin, a member of the Open Law Association board of directors, has long advocated for the openness of legal and administrative data – such as legal texts and precedent. Many legal precedents will be made available to the general public, as well as judges of all levels of experience and expertise, and machines too, of course. Thomas Saint-Aubin believes that “robots will help legal practitioners in exploring these ‘data lakes.’” Speed, expense, and court congestion will be reduced, and the law will likely be more consistent because it will be based more on precedents established in past cases.

Computer code: are they realy intelligent?

There are others who believe that artificial intelligence can replace the legal and judicial professions, even though there are many obstacles to overcome. Opinion in contrast believes that the logic of law and code are fundamentally different.

Judges and lawyers need to interpret and possibly even change the law; existing software is unable to do this. They are frequently able to replicate the past or undertake predetermined activities. Scholars also believe that a machine in place of human will not able to judge the changing requirement of the society and law would stuck in time.

A study conducted in the United States contrasted the outcomes of judges’ judgements with computer algorithms when determining whether an accused should be placed under home arrest or imprisoned pending trial. According to initial impressions, the software is better at predicting how an accused person will act and can thus help make more “objective” judgements. A more broad argument in favor of automating justice is the assumption that computers can be more impartial than humans. However accoutring to Ege Gurdeniz, (dean of AI academy and principal, Digital), Artificial Intelligence (AI) is taught on data supplied by people, and humans are biased. Furthermore, many organizations have a history of discrimination based on their past actions.

Assume for a moment that you’re developing an AI system that will be used to evaluate mortgage applications and provide lending recommendations. In the extreme instance when you have a history of gender bias in loan approval and as a result 90% of women’s applications are refused, the computer will learn this and operate in a biased manner. The situation will deteriorate further.

A human may change in the future, or a circumstance may change as a result of a change in person. However, the machine will continue to act in the same manner as before they were trained.

Algorithms can repeat and amplify human prejudices, in part because they rely on subjective decisions and don’t make choices for themselves, contrary to popular perception. To illustrate, a basic Internet search returns the most frequently consulted results, regardless of how skewed those results may be.

Notary deeds and  blockchain

As part of a contract, as a general rule, two parties undertake to respect a set of obligations: exchange a sum of money for a property, live together by sharing the assets, administer a company in partnership, etc. Will digital innovations (such as the so-called “blockchain” technology) make it possible to do without the legal or judicial actors present (more or less depending on the type of contract) to ensure that the agreement made complies with the law, verify that it is well respected and intervene in the event of a breach – like notaries and judges for a marriage contract or a real estate sales agreement? As Primavera De Filippi explains, “the blockchain works like a decentralized cadastre: it is a way of recording information and certifying this registration at the source, without necessarily going through a third party”. Once constituted in “blocks”, these data make it possible to prove who did what and when, and are in principle impossible to modify or delete. “An interesting application of this principle is the development of ‘smart contracts’,” continues the researcher. This consists of computer-coded the performance of obligations and automating them. Let’s imagine that two individuals agree to enter into a contract: the blockchain allows them to reconstruct the procedure step by step (to prove that it complies with the law) and the automatic execution of the clauses makes it impossible for them to fail. “Even if the parties do not trust each other, they have the guarantee that the actions will be carried out automatically”, summarizes Primavera De Filippi.

💡The idea is all well and good, but the reality is a little more complicated. “The notary is also there to check that the parties are agreeing when they sign,” argues the researcher. Although the blockchain could now support a certification or notarization procedure, it cannot in any way substitute for this process. In contrast, “if the agreement must be broken or modified for unanticipated reasons or circumstances of force majeure, one cannot return and the acts continue to be executed.” For example, automatic transfers to creditors would continue for a corporation that had fallen behind on its payments. However, Primavera De Filippi admits, “We might of course anticipate such scenarios, but it seems difficult to predict everything.” It’s important for the rule of law to have more discretion when dealing with traditional contracts because their natural language is more ambiguous than a computer code like “if this, then that.” “The true challenge would be to take use of the finest of both worlds,” she says. We may conceive, for example, hybrid contracts, with a portion of automation and certification by blockchain, but consolidated by the subjective and more flexible evaluation of professionals.

What is my fear? 😱

I don’t know, why I used the word fear. To make my point clear to you I will give you one real example and one hypothesis.

Amazon stated in April 2019 that a computer system not only monitors warehouse employees’ work but also immediately terminates them if they fail to reach performance targets on a regular basis. The number of packed parcels and absence from work or interruption of work, e.g., due to breaks or restroom slacks, the so-called “time off task,” are the major indicators of performance (abbreviated: TOT). When TOT timings are too long, for example, the system automatically sends out notifications, and the 5% of the workforce with the worst performance statistics is instantly enrolled in a training program – or fired if they consistently fail their targets.

Now imagine a scenario in which big data, the internet of things, and artificial intelligence are used together to monitor everyone all the time. A small proportion of the population is in charge of society, and they see the broader people as nothing more than a machine.