Migle Laukyte, lecturer in the Department of Law at UPF and member of the Evolution of Institutions Observatory Research Group

Although artificial intelligence (AI) and law are two very distant worlds, the distance between them does not prevent them from seeing in each other an enormous stimulus, a challenge, an invitation to understand the limits of creativity and human ingenuity. For years, we have observed the growing interaction and collaboration between lawyers and AI developers. In today’s data-driven society in particular, these interactions have become more intense and prolific. To better understand them, we can divide the relationship between AI and law into two broad areas: legal AI and AI law.

Legal AI refers to the AI-based technological possibilities that facilitate and expedite certain common tasks involved in legal work. For example, expert systems can give advice on how to classify facts; analytical systems can analyse and compare different contracts; and predictive software can forecast a hypothetical trial outcome. The construction and development of these systems is based on knowledge and understanding of how the law works. And therein lies the greatest challenge: understanding the law – in all its varied and complex uses, customs, rules and procedures – and transferring this understanding to programming languages without sacrificing its spirit.

And therein lies the greatest challenge: understanding the law – in all its varied and complex uses, customs, rules and procedures – and transferring this understanding to programming languages without sacrificing its spirit

On the other hand, AI law encompasses all the legal issues related to AI and its impact on society. This gives rise to a world of possibilities: from intellectual property law to human rights; from contracts to the philosophy of law (can a robot be a legal person?); from data privacy to the impact of algorithms on labour rights; from humanitarian law in the case of autonomous weapons to the use of AI in procedural law; from consumer protection to liability for damage caused by machines and smart programmes; and so on. In this regard, lawyers must venture into the world of AI to understand its characteristics and how it functions and, in so doing, its legal impact: explaining how an algorithm based on deep-learning works to a lawyer is no mean feat!

lawyers must venture into the world of AI to understand its characteristics and how it functions and, in so doing, its legal impact

The law and AI thus have many issues in common, and the differences – in methods, objectives, fields and, of course, the time it takes them to respond to changes and new developments – are not only what separates them, but also what draws them to each other, creating reciprocity between them. The important thing is that, whilst AI seeks to reproduce and enhance our capabilities, augmenting our skills and offering us new ways of seeing the world, law guarantees that, whatever AI-based approach or application we might develop tomorrow or a hundred years from now, human beings will always be the end, not the means.