Being a lawyer – and a judge – at the time of ChatGPT

This text is part of the Special AI of Italian Techentitled “Talk to me. The artificial intelligence revolution”, on newsstands free of charge with the newspaper The Republic starting Thursday 23 March

No one would have believed me if, when I entered a law firm for forensic practice for the first time 23 years ago, I had said that the process would be electronic and that registered letters and fountain pen signatures would be replaced with certified emails and digital signatures . Instead, in just a few years, lawyers and magistrates – in some cases forced by the law – have completely changed their working tools.
The revolution is only just beginning though. The advent of ChatGPT (and all the other generative AI solutions) is the beginning of an even more profound change for lawyers and, in general, for the justice system. A sudden change that allows, indeed requires, all legal professionals to rethink their work (and their future). With these tools, anyone can ask any question and, in seconds, get an answer: the text of an email, a sonnet, a lease agreement or a judicial document. The result is often surprising: ChatGPT is the most impressive system I’ve seen in over twenty years of experimenting with technology in the legal field. The uses in the legal field are many: already today these technologies can generate drafts of useful contracts and parts of judicial documents.
Cases of use in the legal sector are multiplying all over the world. In Colombia, ChatGPT was used in drafting a ruling. Some US academics tested it on law school exams: the test passed and the AI ​​wrote legal essays on different topics. Also in the US there are those who have brought a “robot lawyer” to court to decide the defense strategy.
Obviously, there is no shortage of criticism from those who argue that these solutions still make too many mistakes. True, but don’t forget that this is an experimental version. In short, AI can do nothing but improve. Quickly. Very soon for activities such as the first draft of a document or the comparison of deeds and sentences, ChatGPT will be better than a trainee lawyer.
This opens up at least three scenarios.
The first is related to legal services clients. Today, for those who don’t trust Google searches or don’t have a “cousin” lawyer, there is no alternative to turning to a professional. But what will happen when – soon, now – customers will be able to turn to AI systems to write contracts, answer questions and give opinions? At the moment, ChatGPT seems trained to always recommend consulting a professional (also given its experimental nature), but it is easy to imagine that specific services will multiply and will certainly be more convenient than traditional consultancy. To prevent these services from harming users by giving wrong advice, they will have to be developed involving (human) professionals in the sector and will have to provide guarantees in terms of liability in case of damage caused by wrong advice.
The second scenario is related to use in law firms. In recent weeks, Microsoft has announced the integration of GPT (technology at the base of ChatGPT) with Word, the most used program in law firms to write documents. This means that soon part of the texts will be generated directly by the AI ​​on input from the lawyer. But it is likely that artificial intelligence will revolutionize the way in which the research of the jurisprudence underlying the drafting of deeds and opinions is done: no longer for sentences but for defensive theses. Furthermore, AI solutions could be used to allow lawyers, based on court rulings, to know the probability of success of a judicial action, in order to provide the client with transparent information on the risks of litigation.
The third way in which artificial intelligence could revolutionize the world of law is related to the work of judges. There are already several trials of AI solutions that allow judges to write the reasons for sentences, but this is one of the most critical applications of all. AI solutions work on precedents, from which they learn. This means that the sentence written by an AI will hardly be innovative, the “artificial” sentences will be repetitive of what was stated in past rulings. Instead, the world of law is also made up of “revolutionary” sentences that change established interpretations and often represent an advance for rights and freedoms, adapting the interpretation of rules dictated in a different historical context to the times and needs of society. Artificial intelligence can simplify the life of those who work in the justice system, but it can never replace the intuitions of women and men committed to defending rights and freedoms.