Artificial Intelligence in the World of Law

1. Introduction and Context 

The question of what “artificial intelligence” is can be divided into two parts: “What is the essence  of artificial intelligence” and “What is intelligence”. The first question is relatively easy to  resolve, but it points to the question of what can be built. The second question is much more  difficult, raising questions about consciousness, identity, and the mind (including the unconscious  mind), as well as the question of which components are involved in the only type of intelligence  that is generally considered to fall within the scope of intelligence. The study of animals and  artificial systems (not simple models) is beginning to be considered a research agenda in the area  of intelligence. When conceptualizing artificial intelligence, it is assumed that interaction with  the environment faces real needs such as relationships between similar individuals, disputes  between different individuals, pursuit and escape; in addition to causal relationships at different  levels of intuitive understanding (conscious or unconscious) in addition to specific symbolic  exchanges. Talking about artificial intelligence in law is a good way to start correcting common  misconceptions about the use of technology in services markets. For better understanding, a quick contextualization is necessary. 

Previous industrial revolutions mainly served to scale up the mass production of consumer goods.  Innovations made industries more productive and efficient as human manual labour was replaced  by machines. 

Steam machines first appeared, later those powered by petroleum derivatives and electrical  energy. The Third Industrial Revolution brought us computing and automation, increasingly  eliminating the role of the operator. 

Now, we live in the Fourth Industrial Revolution, which is the era of technologies such as artificial  intelligence and machine learning. 

With human work becoming less and less relevant in industries, we have entered the post industrial era, with the services sector concentrating an increasingly larger portion of the world's  workforce. 

And here comes one of the common misconceptions about artificial intelligence: that it is only  applied to industrial production machines.

The increasing sophistication of algorithms allows computers to also perform work that is not  exclusively manual, such as research, reviews, etc. 

Does this mean that Artificial Intelligence will replace all types of work? 

This is the second main misconception about artificial intelligence. What happens is that human  work will have to be, in the broadest sense, increasingly human. 

Including in Law. 

Lawyers, judges, prosecutors, and other professionals in the sector provide an essentially  intellectual service. Artificial intelligence in Law has been proving that there are tasks that can be  assumed and performed by computers. 

The quality level of technology using artificial intelligence has increased significantly in the last  year, with chatbots capable of holding fluent conversations with humans, writing songs or articles,  and generating computer code. Still, experts say that even the most advanced AI systems are no  match for human writing capabilities: They lack a personal voice and style and often produce  wrong, meaningless, or biased answers. In March 2023, a photo that went viral on the Internet  showed Pope Francis parading down the street in an oversized white coat that looked like he had  just come from a fashion show in Paris. The internet went crazy. Eventually, it was realized that  the photo was fake and had been created using a Midjourney AI image generator after a tip from  a Chicago resident. This is just an example of the increasingly complex types of products that  artificial intelligence can produce. 

"I think we're in a time of a lot of questions, and some of them feel quite deep and existential,  especially when it comes to how we think about creativity in an era where artificial intelligence  is going to take center stage in a lot of fields," says John Palfrey, visiting professor at Harvard  Law School, and president of the John D. and Catherine T. MacArthur Foundation.

2. What do the numbers and statistics tell us? 

According to data from the most recent “Future of Jobs” survey prepared by the World Economic  Forum (WEF), by 2027 it is estimated that 69 million jobs will be created and 83 million  eliminated, which will result in a net decrease of 2 % of current employment. Meaning that 14  million jobs will disappear in the next five years or so. 

Another recent study by Goldman Sachs, cited by the Financial Times, concludes that around 300  million jobs in the most developed countries could be at risk. According to the study, lawyers and  administrative staff are among those most at risk of becoming redundant. 

In the United States alone, around two-thirds of workers will be exposed to some level of artificial  intelligence. Many of these workers will see less than half of their workload automated and will  likely continue to work, dedicating part of their time to more productive activities. The study  estimates that this will happen to 63% of the North American workforce, but 7% of that workforce  will be completely replaceable. The impact in Europe should be similar. Furthermore, the study  adds that AI content creation systems like ChatGPT can trigger productivity gains that will  ultimately increase global annual gross domestic product (GDP) by 7% within a decade. 

While some see the emergence of AI with enthusiasm and expectancy, others invoke fears and  dangers that this new technology could bring to Humanity. 

In March, more than 1,000 technology experts concerned about AI's potential to cause harm to  society signed an open letter calling for a temporary halt to future AI development. "Powerful AI  systems should be developed only when we are confident that their effects will be positive and  their risks will be manageable," the letter states. 

Just over a month later, a computer scientist known as the "Godfather of AI," Geoffrey Hinton,  shook the technology world by quitting his job at Google, citing his concerns about the dangers  of AI technology. "It's hard to see how you can prevent bad actors from using this technology for  malicious purposes," Hinton told The New York Times. 

What is certain for now is that these technologies will bring drastic changes to the concept of how  we view jobs and labour production in practically all professions in the world.

3. Two Opinions and Professional Perspectives 

Focusing on the world of Law, a characteristic of any type of revolution is that, while it is  happening, it is difficult to recognize it as such. 

In order to obtain specialized and experienced opinions, I had the opportunity to talk to 2  professionals from the main professions in the World of Law: Law and Judiciary. 

As such, I had the opportunity to speak with Dr. Professor Pedro Caetano Nunes, professor at  Nova School of Law, coordinator of the Master in Law and Management, offered in conjunction  with Nova School of Business and Economics and partner in the area of Dispute Resolution of  the PLMJ Law Firm. Pedro Caetano Nunes was also a judge for more than a decade. 

As well as the Prosecutor of the Republic Dr. Cristina Maria Vieira Alves Janeiro. 

For lawyers, the day-to-day work is likely to change based on what we've highlighted so far. AI  can help optimize much of their work, reserving more time for strategic activities and closer  customer service. It is believed that repetitive, tiring and time-consuming tasks can be easily  resolved and more time can be dedicated to work, leading to greater motivation. 

According to professor and lawyer Pedro Caetano Nunes “Legal professions have some practices  that can become quite repetitive, such as litigation and drafting contracts that imply systematic  action and the emergence of AI can contribute in a certain way to faster and more efficient  performance with possible cost reduction.” 

But will this mean the end of the legal profession? 

Although the odds are greater than changing judges, the fact is that when legal software emerged  (in the 1980s) and even more basic tools, such as document editing programs, many professionals were afraid. According to Prosecutor Cristina Janeiro, “it is unquestionable that AI can help  professionals in terms of jurisprudence research (for example), but I do not believe that they will  completely replace these professionals”. 

Law is not just about cold and objective analyses of the law. Ethical issues must be observed and  the subjectivity of each case must be interpreted. A reasoning that machines are not capable of  carrying out. What will happen is that technology will only replace part of their job.

According to Professor and Lawyer Pedro Caetano Nunes, “In transactional law, many of the soft  skills that constitute a lawyer come from the negotiation capacity that only a human being can  produce. If AI cannot perform these tasks, I believe it also does not have the capacity to replace  a lawyer for now.” 

4. When we talk about technologies in the world of Law, Lawtechs matters 

Lawtech is a startup with a technological matrix that develops products or provides services in  the legal area. The world's leading legal technologies seek to leverage artificial intelligence  solutions in the legal field. This means that all legal technologies develop AI products. There are  also platforms for automation, document management, compliance, communications, arbitration,  legal administration, public data monitoring, professional networking, online conflict resolution  and other types of legal technology. Building a successful legal technology practice requires  major investments in technology. Of course, it is also necessary to know the details of the legal  environment and opportunities for improvement. 

With these Law techs, the automation of some tasks is expected through, for example, RPA  (Robotic Process Automation) technology. This technology is based on the use of robots capable  of capturing data and interacting with programs, websites and systems, carrying out different  types of tasks with the advantages of operating 24 hours a day and with a margin of error close to  zero. The use of these systems will allow a considerable reduction in costs, while achieving  greater productivity and profitability, along with greater effectiveness of the services provided. 

In the judicial area, researchers at the University of Chicago have developed AI technology that  predicts crimes with 80-90% accuracy. How? An algorithm is used that works based on the crime  history of a specific city. Therefore, the system analyses the type of crime, place of occurrence,  date and time, etc. Based on the records, this type of AI then uses that knowledge to generate time  series and predict where and when crimes occur most frequently. 

The negative impact in the criminal area can include deepfakes, driverless vehicles or drones used  as weapons, or forms of cyberattacks, improper access to personal data, scams with manipulation of facial and voice movements, dissemination of false information for defamation or slander or  even new forms of manipulation of the securities market. 

5. More Questions than Answers 

In fact, there are many doubts, if in a judgment phase a fair solution can be reached by a machine,  unlike human beings who are endowed with intuition, emotion and equity, which are elements  that are very important in the decision-making process and to reach a fair decision. Prosecutor  Cristina Janeiro adds “Responses must have a human component, something that machines cannot  perform and all these vicissitudes are important to assign a case with the appropriate prosecution  and decision. AI follows reasoning based on programming but in the course of the case  circumstances and cases appear to us that can alter our decision-making, i consider that AI cannot  reformulate a thought against which it was programmed and as such does not accept different  positions”. 

However, even if we seek to improve these mechanisms, the dangers through the use of algorithms  to detect behaviour, with possible violation of fundamental rights and the violation of the  presumption of innocence, whether through research in computer environments or tools for facial  recognition or voice or to anticipate future conduct. The ease with which this type of approach  can generate situations of discrimination and violation of fundamental rights is immense. In fact,  such concerns motivated a resolution by the European Parliament, in 2021, on the use by police and judicial authorities of artificial intelligence tools, with several appeals to which, the  Portuguese legislator has not yet responded. 

Even if such a result is effectively achieved, would we trust what is the most cherished and  fundamental human right, which is freedom, to be dictated by a mere machine powered by  artificial intelligence?

From my perspective, I don't believe that it will become a reality in the near future, it could help,  but not replace. AI will allow lawyers to reduce the time spent analysing both evidence and  studying the underlying law and will provide greater legal rigor when making decisions. 

A robot will not be able to replace a judge and hand down sentences based on algorithms, since  the rule in criminal proceedings is that the evidence is freely assessed by the judge, in accordance  with the rules of common experience and what he learns through hearing witnesses. However,  law and justice cannot ignore the fact that AI can become one of its most important assistants or  adjuncts, whether for organizing procedural documents, for processing the evidence in the process  itself, or even for research into jurisprudence and doctrine, contributing to more agile processes  and increasingly faster decisions. 

What is certain is that robots are starting to take their first steps autonomously. In February, the  world's first robot lawyer made his court debut and defended his first client. This was the first  case in which AI took the place of a Human Being in defending a case in court. The robot, created  by startup DoNotPay, gave instructions to the defendant on the best way to argue against a  speeding ticket. 

But how were the instructions given? Easy, the robot was present on a cell phone and sent  warnings through a headset to the defendant, who agreed that he would only say what the AI told  him. This was a pilot case that served to test DoNotPay's services.

6. Stopping evolution or Regulating Artificial Intelligence? 

Everything will depend on how AI evolves in the future, but it could undoubtedly push the limits  of existing law. 

On September 21, 2023, we were surprised by the news that ChatGPT was trained and managed  to write the next book by the author of “Game of Thrones”, with George RR Martin suing OpenAI  in which a final legal battle against theft of Copyright and intellectual property rights. Part of the  existing legislation that may need reformulation and involves the definition of a derivative work.  Current copyright regulations state that "to be copyrightable, a derivative work must incorporate  part or all of a preexisting work and add new original copyrightable authorship to that work.  There's still a lot of interpretation around the edges that the courts must do. And with advances in  AI, these interpretations become “more complicated”. 

Regarding Regulation and determination of Responsibilities, Prosecutor Cristina Janeiro  concludes that “In a specific case, determining who has control of the fact seems difficult to  determine. Everything will depend on the way the legislation is created and only from this  legislation will it be possible to apply it to the case. To defend the Human Being, it is imperative  to resort to Collective Responsibility, that is, entities that represent AI (which can be several  companies at the same time) to determine responsibilities. Legislating use is essential because  without it, it’s not possible to attribute conduct to AI. Once legislation is created, it may be in the  best public interest to establish a new regulatory entity that ensures and monitors its use to establish and protect the limits imposed by law.”

7. What we should keep in mind 

We easily conclude that as Artificial Intelligence contributes to the performance of the professions of law and judiciary, we are also able to carry out a logical-deductive exercise that here too there  is an opportunity for Law and access to justice to cease to be seen and labelled by many as a  privilege or elitist access. What is certain is that it would be easy to understand that we could use  reasoning that a lawyer using technology and using artificial intelligence can be seen as an  additional cost for the professional in question and that as such would lead to an increase in the  price of fees. However, as we previously analysed, it is also expected that the time spent by the  professional will be optimized to perform certain tasks that could be more time-consuming and  as such they can save this time that they would otherwise end up spending and focus on more  cases, and on other activities. That will allow them to make their activity even more profitable.  As such, with the professional able to monetize their work, we apply one of the main economic  laws to the specific case, which is the law of demand and supply. By increasing supply, demand  will decrease as well as prices so that they are more competitive and appealing to customers.  Naturally, justice should not be seen as a business, but rather as a fundamental right to which all  Human Beings must be entitled and Artificial Intelligence appears to us as a tool that gives us the  possibility of achieving this objective.