Press Releases

“Law & Technology: The Future of Law and Justice in the Age of Artificial Intelligence” 

First of all, I would like to thank the Cyprus Bar Association and the European Legal Tech Association for inviting me to the 7^th Educational Conference of the Cyprus Bar Association.   

In the late eighteenth-century technological developments such as the steam engine “enabled unimaginable physical feats, leading to mass production, mass transportation, and railways, transforming therefore the face of the world.” This was termed as the first machine age.

Veering towards the so called “second machine age” academics argued that new computer technologies are leading to a transformation equivalent to the first machine age. Namely, it was stated that “computers and digital advances are doing for mental power what the steam engine and its descendants did for muscle power.”

This is where Artificial Intelligence (AI) comes in. What is AI? AI might be described as using technology to automate tasks that “normally require human intelligence”. It must be noted that no single definition of artificial intelligence is accepted by the scientific community.

However, the European Commission, via article 3 of the draft Artificial Intelligence Act, sought to establish a legal definition of the term “artificial intelligence system.”

Namely, it means “...software that is developed with [specific] techniques and approaches [listed in Annex 1] and can, for a given set of human-defined objectives, generate outputs such as content, predictions, recommendations, or decisions influencing the environments they interact with.”

According to Annex 1 of the draft Act, techniques and approaches used today to develop AI include “machine learning”, “logic and knowledge-based” systems and “statistical approaches.”

AI in Law involves the application of the aforementioned techniques “to make law more understandable, manageable, useful, accessible or predictable.”

It may be useful to divide the use of AI in law, on the one hand, by those who apply the law (judges) and on the other hand by those that use AI in legal practice (lawyers, paralegals, researchers etc.). Finally, allow me to add the use of AI in the context of legislative drafting given my role as the Law Commissioner of the Republic of Cyprus.

AI used in legal practice may take several forms. Amongst others these include E-Discovery, Legal Research, Expertise Automation, Document Management and Predictive Analytics.

E-discovery is considered the most established use of AI in law indicating to software that analyses a large volume of documents identifying those which are relevant to the search criteria, in less time, for a less cost and with far more accuracy than a number of humans and more specifically a number of lawyers.

Warning about the limits of such software, we are reminded that lawyers are those who have the final say as to which documents are relevant or not to the law. This is because e-discovery, and I quote, “cannot make those decisions which involve understanding the law and the facts and dealing with strategy, policy, and other abstractions that AΙ technology today is not good at dealing with”.

We now turn to the use of AΙ by the judiciary. Besides the use of similar software with those used by the legal practitioners aiming to accelerate legal research, the judiciary appear to have different uses for artificial intelligence. Namely, the courts in the United States have been seen to use risk assessment tools in the criminal justice system. An example of such tools is the use of software that employs AΙ by judges, to determine (through a percentage) a defendant’s risk of reoffending. In order to make such prediction the algorithms behind such systems use past crime data.

This, must be noted, contains an underlying danger. Namely, it was stated that “computed models that learned pattern from data may be subtly biased against certain groups of people based upon biases embedded in that data”. A hypothetical example that was given is where “biased police activity may be embedded in the recorded police arrest data and thus the machine-learning system that relies on such data may encode the related biases” targeting, therefore, a certain group of persons.  Such bias is explicitly dangerous where the judiciary uses AI in order to determine a defendant’s risk of reoffending as was seen above.

At this point, let me explain how Artificial Intelligence can contribute in the legislative drafting process which is an integral function of the Office of the Law Commissioner of the Republic of Cyprus.

It is not expected that AΙ systems will completely replace a human legislative drafter given that drafting legislation contains too many unpredictable variables, making it impossible for AΙ systems to formulate usable statutory provisions themselves.

However, AΙ could potentially take an active part within the legislative drafting process in various specialised ways.

Firstly, the starting point of the legislative drafting process is for the drafter to understand the policy objective behind a proposed bill. Here, AΙ systems could determine whether the national proposed policy objectives are in accordance with international obligations and best practice. For instance, they can ensure whether the policy objectives are in accordance with European Union Law and International Law in order to further predict whether the proposed national bill will also meet such international standards.

Finally, the legislative drafting process stresses the need for careful proof-reading, because it gives the drafter the opportunity to check for errors in language, such as, grammar or spelling mistakes or whether plain language is used instead of archaic language. Hence, it is stated that the inclusion of automated proof-reading technology developed specifically for the legislative drafting process, could be vital to detect language errors that a human drafter could not initially detect.   

After analysing certain examples of the use of AΙ in both the legal practice and the judiciary, one realises that the use of AΙ within the Law concerns majorly the tasks that are mechanical and repetitive. It is stated, however, that Lawyers “are still required to provide judgment, empathy, creativity and adaptability”, postponing the fear of legal practitioners as to their replacement and their redundancy by AΙ.

Nevertheless, it may still be reasonably argued that the role of the lawyer ought to be redefined. Namely, it was stated “that future lawyers will need to be trained in order to assess the relative strengths and weaknesses of AΙ solutions to meet their clients’ needs.”

A recent example of this is where major global law firms partnered with a company that provides on-demand learning programs for law firms and corporations. These aim to train legal professionals and equip them with the necessary skills to effectively use emerging tools in the field of generative AΙ, ensuring their competitive place in a future legal environment.

In conclusion, after listing the benefits of AΙ, as well as certain concerns as to its development, I believe that we should not view AΙ as a threat and most importantly we should not ignore its existence given its transformative role regarding many aspects of the legal practice, the judiciary and hopefully the legislative drafting process in the years to come. If understood and used effectively and subject to proper regulation, AΙ within the Law, Justice and the Legislative drafting process can truly lead our profession to higher levels in regards to efficiency, precision and the general quality of the Law.

Thank you.

(NG/SCH)