Technological development, global access to the internet, the physical dematerialisation of evidence and the growing importance of the influence of artificial intelligence are key trends today. It is therefore unavoidable to critically study the interaction between artificial intelligence and civil procedure, be it in the processing of the case, in the control and verification of procedural requirements, in the preparation of orders, in the production and evaluation of evidence, and finally in the preparation and drafting of the decision.
The advantages appear to be manifold, multi-faceted and very relevant: think of the increases in speed and efficiency in litigation, the medium-term reduction in the cost of justice and the mitigation of the risk of human error. However, it is important not to overlook the obstacles that arise when the effects of applying AI to the process are placed in the light of values dear to Civil Procedural Law.
One example that immediately springs to mind is the potential shortcomings in the explainability of decisions produced by autonomous systems, which could imply a lack of justification for the judicial decision, in conflict, therefore, with a right of the parties that, in our country, deserves constitutional protection. Predictive justice is also a source of relevant axiological doubts. It is announced as a very useful tool, especially from the perspective of lawyers, as it could make it possible to predict the outcome of a claim depending on the court that will judge it.
However, at this point the risk arises that judges will be conditioned by previous decisions, no longer because of an institutional precedent, but because of a coherence argument, which may well be introduced in closing arguments by lawyers. Once available on the market, such a tool could also be used by judges, with self-conditioning effects equivalent to those mentioned above. Moreover, such predictive tools or similar technologies could even make it possible to build machines capable of autonomously producing a decision. Once the facts have been established, either by agreement or by a human judge, the machine will simply search the database for factually and legally similar cases, adopting the statistically prevailing decision.
This possibility raises interesting questions: starting with the risk of data bias, and ending with the ablation of one of the driving forces behind the evolution of the law, which is the controversy over the best solution to the case. To conclude the list of examples, it should not be forgotten that the development of technology is (and may continue to be) an expensive business, which forces software companies to sell their products at high prices or to many people at a more affordable price.
In Portugal, any market is known to be small in size, and this is confirmed by the trade in products aimed at legal professionals. Therefore, if we accept the need to set a high price for these products, we will probably create information and resources asymmetries, since litigants with greater economic capacity will be able to benefit from AI tools that give them significant advantages, while others will not, which will accentuate pre-existing imbalances.
Diagnosing the risks is fundamental, but insufficient, especially if we accept that the advance of technology and its use is inevitable, as well as the real advantages that can be extracted from it for civil proceedings. Therefore, the next step is to understand and propose ways of preventing or eliminating the risks identified. This completes the course of a reflective but also practically relevant study.
This project is concerned only with civil proceedings. It would naturally be possible to cover other processes in which these issues are relevant, such as criminal proceedings, but the risk of dispersion is high enough to recommend a more restricted scope. What has been said does not rule out the clear susceptibility of generalising the conclusions reached with reference to civil proceedings to other cases.
The legal framework for this project is Portuguese. However, the integration of foreign researchers seems very pertinent, as it will make it possible to learn about, compare and integrate current thinking in other legal systems on identical issues. As this project deals with legal issues permeated by technology, it is essential to have consultants from the fields of computing in general and artificial intelligence in particular.
Decoding the nature and mode of operation of the machines, as well as clarifying doubts about the type of technology under analysis, is essential if the research is to be both rigorous and adherent to reality. The partnership with the Max Planck Institute Luxembourg for International, European and Regulatory Procedural Law is expected to be extremely fruitful, given that it is a research centre of excellence in the area of civil procedure. The partnership involves the participation of a researcher from the MPI Luxembourg in the project, as well as the organisation of research periods for the project's researchers at the MPI Luxembourg.
As there are several sub-questions to be answered, four teams of researchers will be formed, each dealing specifically with one (or possibly two) sub-questions. This split does not imply a breakdown of the research, but rather a continuous collaboration between all the researchers, which will materialise in joint meetings.