PREPRINT
Revista Ítalo-Española de Derecho Procesal
pp. 1-19
Madrid, 2026
DOI: 10.37417/rivitsproc/3349
Marcial Pons Ediciones Jurídicas y Sociales
© Alessia Bastianelli
ISSN: 2605-5244
Recibido: 06/02/2026 | Aceptado: 09/03/2026
Editado bajo licencia Creative Commons Attribution 4.0 International License.
THE FUTURE OF CIVIL JUSTICE IS TODAY: DIGITALISATION ACROSS THE VARIOUS STAGES OF CIVIL PROCEEDINGS IN ITALY
Dr. Alessia Bastianelli*
University of Tuscia
ABSTRACT: This article analyses the digitalisation of Italian civil justice, from the Telematic Process to Artificial Intelligence. It examines the Riforma Cartabia and the National Recovery and Resilience Plan (PNRR) regarding the Ufficio per il Processo and judicial reorganization. Through pilot projects, the study assesses digital technologies in the introductory, preliminary, and decision-making phases. Finally, it evaluates the future prospects and limitations of Artificial Intelligence in civil proceedings.
KEYWORDS: civil proceedings; digital justice; e-justice; judicial case management; judicial decision-making support system.
SUMMARY: 1. INTRODUCTION. FROM DIGITAL JUSTICE TO GIUSTIZIA AGILE.— 2. THE DIGITISATION AND THE USE OF AI IN CIVIL PROCEEDINGS: 2.1 Digitisation in the introductory phase; 2.2 Digitisation in the investigative phase; 2.3 Digitisation in the decision-making phase.— 3. FROM GIUSTIZIA AGILE TO PREDICTIVE JUSTICE. PROSPECTS AND LIMITS IN THE USE OF AI.— 4. BIBLIOGRAPHY
The process of digitising Italian justice is a phenomenon involving the technological implementation of judicial offices, but also a change in the way these offices work. In fact, while the primary objective in the initial phase, which began with the advent of the PCT, was simply to dematerialise document flows, the current drive for reform under the National Recovery and Resilience Plan (hereinafter NRRP) 1 has led to a paradigm shift: from the automation of judicial offices to the overall reorganisation of their workflows.
The transformation of civil justice has taken place over time through a series of steps 2.
Firstly, the concept of transparency and accessibility was redefined. The introduction of the Electronic Civil Procedure (PCT) on 30 June 2014, i.e. a portal dedicated to tracking the judicial events of a case, has allowed defence lawyers and citizens to monitor the status of proceedings in real time, breaking down the physical barriers of court registries and drastically reducing communication downtime. With this advent, an initial transition had to be accepted: from ‘digital document’ to ‘digital data’; information is no longer understood as a simple scanned image, but as structured information that is digitally originated, thus lending itself to being analysed, categorised and processed in an automated manner.
Subsequently, the pandemic emergency acted as a catalyst for the telematic management of hearings, consolidating the use of remote connections and written proceedings. The legislator has gradually introduced a series of exceptional provisions, which have had an impact on the ordinary conduct of civil proceedings 3. These methods have demonstrated the effectiveness of adversarial proceedings not necessarily linked to physical presence, significantly speeding up the interlocutory stages of the proceedings.
Two years later, the Riforma Cartabia, implemented by Legislative Decree 149 of 2022 4, endorsed the measures tested during the emergency phase, laying the foundations for a civil process based on the digital paradigm. The legislator has carried out a complete codification of telematic practices, introducing crucial provisions into the Code of Civil Procedure, such as Articles 127-bis and 127-ter, which regulate, respectively, hearings via audiovisual link and the replacement of hearings with the filing of written notes; the latter of which has since become the standard form for the management of interlocutory hearings and postponements. Furthermore, through the inclusion of Article 196-quater of the implementing provisions of the Code of Civil Procedure, the mandatory nature of electronic filing for all procedural acts and documents was established, extending this requirement to proceedings before the Judge of the Peace and court-appointed technical consultants. Consistently, the new Article 196-quinquies of the Italian Code of Civil Procedure establishes that the court file must be kept exclusively in electronic format, requiring that every court order be digitally signed. This framework has been completed with the strengthening of electronic notifications referred to in Articles 147 and 149-bis of the Italian Code of Civil Procedure, which require defence lawyers to use certified email whenever the recipient can be found in public directories.
In the broader context of the reform of the Italian justice system, the NRRP programme was then forcefully introduced, which led to the strengthening of the Ufficio per il Processo 5: an empowerment that involved not only an increase in staff, but also the creation of a ‘digital operations centre’ at the service of the judiciary, establishing itself as the ‘pivot of the new organisation of justice’ 6. The basic idea was not to limit the reform of the UpP to a reform of rules, but to implement a reform of ‘people and resources’, investing heavily in human capital and digital infrastructure. The UpP has, in fact, taken on the functions of a support unit where young lawyers and economists (the so-called UpP staff) are responsible for managing electronic document flows, as well as performing clerical and support tasks for magistrates, including studying case files, researching legal precedents and preparing draft measures 7.
Subsequently, the regulatory strengthening of the UpP was followed by a programme aimed at extending this institution to judicial offices where this tool was absent or little used. In this sense, digitisation and organisational efficiency have become the pillars of a unified strategy promoted by the Ministry of Justice through the ‘Progetto unitario per la diffusione dell’Ufficio per il Processo e l’implementazione di modelli operativi innovativi negli Uffici giudiziari per lo smaltimento dell’arretrato’ as part of the ‘TASK FORCE per gli uffici per il processo’. This initiative, created as part of the PON Governance e Capacità Istituzionale 2014-2020, has operated in close synergy with the structural interventions provided for in the NRRP. The primary objective of this project was to overcome the systemic critical issues that have historically plagued civil jurisdiction, identified mainly in the excessive length of proceedings and the accumulation of backlogs. Digitalisation has been elevated to a strategic driver for achieving a quantitative and qualitative objective: the elimination of the backlog and the reduction of the average duration of proceedings (disposition time) 8. In line with this action, several projects have been launched by the Directorate-General for the Coordination of Cohesion Policies, in collaboration with 26 Court of Appeal Districts and 56 universities throughout the country. The universities participated in a consortium, divided into 6 macro territorial areas 9, but with a common goal: to spread the UpP with actions aimed both at managing the backlog in judicial offices and preventing its formation; prevention that took place through the training of UpP staff but also through a ‘bottom up’ approach, by adapting university education to the real needs of the justice system, with the aim of training lawyers prepared to face the future challenges of Italian justice.
Before looking at how digitisation and the use of AI systems in civil proceedings 10 , many of which were developed as part of the PON Governance programme, have proven to be able to intervene in a targeted manner at certain procedural junctures that characterise civil proceedings, it is necessary to observe how technology can become a catalyst capable of bringing about an overall improvement in the timing and effectiveness of the judicial response as a whole. Even before acting on specific procedural stages, the implementation of digital models makes it possible to re-engineer the way judicial offices work, reducing overall decision-making times and optimising the allocation of resources.
An initial example of possible digital implementation is represented by a case file weighting system, developed as part of the Giustizia Agile project conducted by Macro-area 3 11, led by the University of Tuscia, and not yet in use in Italian courts. This system should come into operation when cases are registered, allowing the human resource requirements within judicial offices to be objectively quantified from the outset. This calculation is based on a linear formula: the volume of proceedings is multiplied by a difficulty coefficient (the ‘weight’), thus determining the total number of hours required to complete the case. This figure, relative to actual production capacity, by dividing the total hours worked by the product of the time each magistrate can devote to the jurisdiction and the number of judges actually in service, provides an accurate indication of the number of magistrates needed to meet the demand for justice in the period considered. This leads to a mathematical consequence: if the human resources available at a judicial office are fewer than the value obtained from the calculation, there will inevitably be a delay in resolution times and the formation of a new backlog. This means that this calculation would allow for an objective assessment of the overall workload of the judges in the judicial office (and consequently their numerical sufficiency or insufficiency) and also the workload of each individual judge. At present, however, the distribution of the workload of individuals is carried out using only macroscopic parameters, such as subject matter or procedure. The weakness of the current case allocation criterion is that it does not allow the real complexity of individual disputes to be taken into account. The assignment of cases is still based on criteria of automatic rotation and equal distribution, aimed at ensuring an identical number of cases for each magistrate. While this method guarantees numerical balance, it neglects the endogenous variables of each case, resulting in an unequal distribution of work in terms of the judicial effort required. For example, the simultaneous assignment of two cases with the same subject matter in family law to two different judges does not mathematically determine the same use of effort and time: the handling of a judicial separation compared to another separation of the same kind but with custody of minor children involves a radically different expenditure of energy for the magistrate. In this context, the development of AI-based systems would make it possible to overcome these limitations; through the processing of structured data, AI could determine in advance the complexity of each case and assign it to the magistrate who, at that moment, has a role with a lower ‘weight’. The aim is to move from a distribution based on quantity to one based on the actual sustainability of the role of each individual judge.
A further example of the potential impact of new applications in the civil sector, and in particular in the commercial and bankruptcy sectors, was the experiment carried out by the Court of Catania, following the establishment by the Inspectorate of the Ministry of Justice of a technical committee tasked with implementing the Inspectors’ Package for the extraction of data of interest for inspection from the SICID and SIECIC applications. This experiment demonstrated the usefulness of a ‘management dashboard’ in supporting the work of the Section President. In fact, the Sicilian court tested an application capable of providing a snapshot of ongoing proceedings, those most behind schedule, the average duration of trials, the time taken to file measures, data relating to assignments given to professionals, the productivity of judges and the detection of any false pending cases 12. Such a tool can be useful in various ways and to various parties: firstly, for the section president himself, not so much as a means of control over other judges, but as a tool for managing the section; secondly, for individual judges in the section, in order to monitor the fair rotation of professional assignments, such as appointments as bankruptcy trustees or support administrators: assignments that involve significant burdens but are also sometimes very lucrative; thirdly, this tool could prove useful for the overall management of the entire office, acting as an accountability tool for the entire court 13.
Furthermore, with regard to the daily activities of each magistrate, the projects within the NRRP have highlighted that the work of judges could be organised more rationally through the creation of dedicated software. For example, a calendar of hearings could be created as a tool to rationalise postponements and organise the judicial activity of individual judges. In fact, although a trial calendar has been required by law since 2009, current practice tends towards the atomistic scheduling of individual hearings, which ends up postponing the resolution of more complex cases. On the contrary, the use of intelligent programming algorithms would make it possible to define the entire timeline of the proceedings from the moment of registration, satisfying the need for certainty of timing through a system calibrated to the actual times of civil proceedings operating according to a FIFO (First In First Out) logic, in which disputes are dealt with in sequence according to the order of registration 14.
Furthermore, with a view to reorganising the work of magistrates, a turning point could be represented by the use of software designed to perform repetitive calculations or draft standardised reports or procedural measures for the entire proceedings, which are currently entrusted to civil judges. An example of this is the calculation of the amount awarded for maintenance payments in separation and divorce cases. The use of AI, which must necessarily be coordinated with the provisions of the Civil Code and with the function of maintenance payments outlined by the most recent guidelines of the Court of Cassation 15, would respond to the need to relieve the judge of purely computational tasks. The use of dedicated prototypes, such as the one developed by the University of Florence, would make it possible to process data for the calculation of maintenance payments (such as net monthly income, any family or instalment expenses, the presence of children with disabilities and the allocation of the marital home) 16, significantly simplifying the time required to settle family law proceedings. Alternatively, AI could be used to create software integrated with the ministerial tables or those issued by the Milan Civil Justice Observatory 17 for calculating compensation for non-pecuniary damage pursuant to Article 1226 of the Italian Civil Code. It would be useful for magistrates to use such software, and it would be even more useful to access it through the magistrate’s console application 18.
Furthermore, the implementation of specialised applications would prove useful not only for magistrates, but also for the tasks entrusted to UpP officials, who are legally responsible for supporting the judiciary by preparing draft measures. The use of advanced legal research tools by UpP officials would transform the drafting of documents from a manual and repetitive activity to a guided and structured process. AI could be used to draft the most frequently used procedural documents, such as the new parenting plan 19 pursuant to Article 473-bis 12 of the Italian Code of Civil Procedure used in separation or divorce disputes: AI could draft a prototype parenting plan ad hoc for each case, automatically entering data such as the children’s daily commitments and activities, extracurricular activities, eating habits and medical expenses from the case file; or AI could draft and update settlement decrees for legal aid or even court hearing minutes. In these cases, the technological contribution would not be limited to the computational component, but would support the judge or court clerk in the preparation of procedural documents, which, although simpler than writing a ruling, still require a considerable amount of energy and resources.
The specific procedural stages in which digitisation and the use of AI can help to reduce the time taken by the justice system can be identified by comparing the normal procedural timeframes of civil proceedings (and in particular ordinary proceedings) provided for by the Code of Civil Procedure and the abnormal ones, caused by staff shortages or the lack of effective applications in the judicial offices; an analysis which, for example, in the Giustizia Agile project conducted by Macro-area 3, was called ‘mapping’ of proceedings 20 and was carried out by creating flowcharts used to document the flow of a given process.
This activity, with regard to the introductory phase of the trial, demonstrated that digitisation and the implementation of AI would make it possible to transform the electronic file from a mere static container of documents into a dynamic information tool. Specifically, the application of algorithmic technologies in the timeframe between registration and the new preliminary checks referred to in Article 171-bis of the Italian Code of Civil Procedure 21 can enable the generation of a ‘file identity card’. This tool, designed to evolve with each procedural event—such as the emergence of necessary co-defendants pursuant to Article 102 of the Italian Code of Civil Procedure or due to third-party proceedings pursuant to Article 106 of the Italian Code of Civil Procedure—extracts and summarises the identifying elements of the claim, such as the parties, the monies sought and the value of the dispute, providing the magistrate with an immediate and constantly updated overview of the case. This information sheet, in fact, could differ from the mere listing of “events” present in the application already used by judges, the Consolle del Magistrato, proving to be a useful tool for planning the trial calendar and for an initial prognosis of the case. It could also serve as a digital historical record: it could be a document that can be consulted at every stage of the trial, aggregating the history of the case, the magistrate’s notes, the studies carried out by the UpP staff and the critical issues identified, accompanying the judge and his assistants towards an ‘assisted decision’. In addition, this case file identity card could include a comparison of the different arguments put forward by the parties, extrapolated in turn from software capable of comparing the different reconstructions of the facts prepared by rival lawyers 22.
A further aspect that could be identified through the use of AI systems in the introductory phase of the trial, and which could be detected even more effectively thanks to the drafting of a case file summary, is the mediability index of the dispute 23.
Starting from the assumption that litigation should always be the last resort in conflict resolution 24, the creation of an indicator capable of assessing in advance the suitability of a dispute for Alternative Dispute Resolution (ADR) and, in particular, for mediation 25, appears to be a pioneering objective. This approach aims to reduce the abnormal recourse to trial, in accordance with the principles of efficiency and reasonable duration enshrined in Article 111, paragraph 2, of the Constitution. This objective was also pursued with the Riforma Cartabia, which gave mediation an unprecedented central role, strengthening the regime of Legislative Decree No. 28 of 4 March 2010. In light of the new regulatory landscape and building on the experiments already underway in Italy prior to the NRRP, the Giustizia agile project has continued its studies on the mediability index of disputes in civil and commercial cases. The aim of the research was to identify a tool capable of highlighting not only the existence of a condition of admissibility that has not yet been fulfilled, i.e. the need to go through the mandatory mediation procedure pursuant to Article 5 of Legislative Decree No. 28 of 2010, but also to estimate the parties’ propensity to settle in matters where mediation is optional and thus assess the possibility of resorting to referred mediation. For the latter, the judge must analyse, pursuant to Article 5-quater, paragraph 1, of Legislative Decree No. 28 of 2010, the nature of the case (the case must in fact concern one of the matters provided for by the law, such as relations between neighbours, family or corporations), the behaviour of the parties (and therefore, for example, their willingness, even partial, to mediate), the status of the case (referral to mandatory mediation is possible for proceedings already pending on 28 February 2023 at the hearing under Article 183 of the Italian Code of Civil Procedure or after the taking of evidence; for proceedings initiated after the Riforma Cartabia. However, referral to mediation is also possible during the introductory phase of the proceedings, at the time of the preliminary checks pursuant to Article 171-bis of the Italian Code of Civil Procedure and in any case until the case is referred for decision) and ‘any other circumstances’. Following these parameters and using the practical tools created so far for the use of this operating model (i.e. the dispute form containing the factual and legal reconstruction of the case, the draft order for referral to mediation, the list containing the cases that may arise, such as participation or non-participation in mediation, and the Excel spreadsheet template for returns), it is possible to identify the dispute’s mediability index 26. This acts as an early screening tool: by identifying the characteristics of the matter in dispute and the history of the parties, this index, if obtained, can suggest to the judge and the UpP staff whether referral to mediation is appropriate. The advantage of the mediability index is clear: it optimises the use of judicial resources and promotes a faster, less costly resolution of social conflict that is more oriented towards the reconciliation of different interests. So much so that in the courts where it has been tested, it has yielded excellent results 27; one could predict even better results if this operating model could also make use of AI in the near future.
The area in which AI systems appear to be least developed is the preliminary investigation phase, i.e. the phase of civil proceedings that ranges from the briefs pursuant to Article 171-bis of the Italian Code of Civil Procedure to the hearing pursuant to Article 183 of the Italian Code of Civil Procedure. The preliminary investigation phase currently appears to be the least amenable to the use of AI, as it is based on qualitative and relational assessments that are difficult to automate without affecting the guarantees of a fair trial 28. As a result, the statistical-probabilistic nature of AI systems encounters structural limitations with respect to principles such as the assessment of evidence 29. AI cannot intervene in the ascertainment of facts 30 because it is the judge who must make a discretionary assessment of the cognitive effectiveness of evidence whose value is not strictly predetermined. So much so that even Article 15 of the new law no. 132 of 2025, which came into force at the same time as the adoption of the AI Act 31, reiterates this assumption. In fact, in order to limit the use of this tool in judicial proceedings, the law expressly states that any decision ‘on the assessment of facts and evidence’ is always reserved to the magistrate.
Yet AI systems could facilitate, if not the assessment of evidence, at least its formation or the categorisation of the evidence gathered, especially in certain matters subject to litigation.
AI could be helpful in activities leading up to the formation of evidence, such as choosing a court-appointed technical consultant 32. This could contribute to the selection of the most suitable consultant, where suitability could be ensured by combining the highest level of expertise (to be examined through an analysis of the professional’s qualifications and documented experience) with the principle of rotation in appointments, in order to minimise the risks involved in choosing consultants. In addition, with reference to the same means of evidence, AI could also be used to formulate questions for the preparation of the expert report or to evaluate the outcome of the consultancy at the end of the mandate 33. Furthermore, AI could be used in witness testimony, first during the admission of evidence 34 to verify the relevance and consistency between the evidence presented by the party and the subject matter of the dispute, and then to assess the reliability of the witness. In the context of documentary evidence, on the other hand, it could be used to identify the different types of documents (e.g. public deeds or private deeds) and also to verify their authorship 35.
Furthermore, in specific proceedings, such as those in the field of immigration law, AI could support judges in carrying out various investigative activities. For example, in the pilot project led by the University of Catania as part of the Giustizia Smart: Strumenti e modelli per ottimizzare il lavoro dei giudici (JustSmart) project carried out by macro-area 6, AI has proven to be a practical tool for organising preliminary investigations in international protection cases. An application has been created for this sector that allows information to be extracted regarding the outcome of previous investigations that have already been decided, in order to support the judge in another investigation. Thanks to the file matching function, which allows users to select from predefined fields those that reflect the migrant’s situation in the specific case (e.g. country of origin, countries of transit, language, specific vulnerability), this application allows users to extract an ordered list of measures relating to ‘similar’ investigations, ranked in terms of correspondence to the case in question using a similarity score 36. Furthermore, in the same sector, AI could assist judges in quickly finding Country of Origin Information (COI), facilitating the exercise of the unofficial investigative power that the judge assumes in this type of proceeding 37. By entering a few keywords, the judge could learn about the current situation in the applicant’s country of origin in a matter of seconds, provided that these databases are first created and then continuously updated.
With regard to the cataloguing of evidence, however, the use of AI could extend not only to the organisation of the evidence submitted by the parties, but even to the identification of only the evidence that is actually relevant to the decision, through the use of specific machine learning algorithms. Starting from the subject matter of the dispute, AI could be used to select documents relevant to the decision, excluding those that are irrelevant, given that, according to some 38, in most litigation, the relevant sources of evidence are already predetermined and limited on the basis of regulatory parameters, established interpretative criteria and procedural rules. However, such a perspective raises significant critical issues. In particular, there is the problem of identifying who is responsible for verifying, on a case-by-case basis, the correctness of the criteria adopted by the algorithm in selecting and categorising evidence for each case. Should judges limit themselves to checking the correctness of the cataloguing or are they also required to verify the correctness of the criteria used by the machine? Considering that ordinary judges do not, as a rule, have the technical expertise necessary to evaluate the functioning of complex algorithmic systems, it would seem necessary to introduce a technical support figure at this stage, similar to a court-appointed technical consultant, tasked with verifying the reliability of the selection criteria adopted by the AI 39. The use of AI would therefore be useful for a more rapid selection of relevant evidence, but the need for an additional technical consultant would lead to a paradoxical effect: instead of simplifying the preliminary investigation phase, the use of AI in this case would risk adding a further element of complexity.
Finally, the use of AI in the preliminary investigation phase could also be applied in relation to operational needs that arise during this phase, in particular with regard to the automated transcription of hearings 40. The use of voice recognition systems would make it possible to obtain, in a short time, a complete and structured transcription of the statements made in court, the questions asked by the judge and the observations of the parties, as is the case in criminal hearings. It could even be used to translate these statements. If used as mere technical aids, these tools would not directly affect the assessment of the evidence, but could help to strengthen the traceability of the investigation and the quality of the reasoning behind the judgment, providing the judge with a complete textual record of the proceedings, in addition to the minutes.
In the decision-making phase of civil proceedings 41, AI and legal analytics tools could primarily contribute to the drafting of judicial decisions. AI systems, currently being implemented, have proven capable of supporting judges in reconstructing the factual and legal aspects of a decision, identifying relevant regulatory provisions, and preparing clear and concise final decisions 42. In this context, legal writing assistant projects such as Cicero 43 developed by La Sapienza University are a concrete example of applications designed to contribute to the construction of the jurisprudence of the future 44. Fine-tuned on approximately 20,000 civil judgments selected from a corpus of over one million decisions provided directly by the Ministry of Justice, the Cicero application has been programmed to enable the rapid and systematic integration of regulatory and jurisprudential references, offering standardised draft measures that replicate the style of the source documents and thus the style of the magistrates in order to facilitate their work. If one of the main critical issues of the justice system lies in the sluggishness caused by the many tasks weighing on magistrates, applications such as Cicero could make a difference. Although the judge would still have to intervene to verify the adherence of the general statements to the specific case 45, the drafting of the entire measure is one of the most time-consuming phases for the magistrate. As a result, this tool could provide concrete, albeit partial, assistance to Italian courts. This is also because Cicero’s usability is not limited to magistrates alone, as it can be widely used by court clerks.
Secondly, following the adoption of the decision, AI can intervene in five main areas that are relevant for subsequent decisions.
Firstly, AI can be used to collect measures, aggregating judgements and orders to enable systematic analysis of rulings. Secondly, it can be used to ‘chew’ data through processes of anonymisation and maximisation of content (also under the supervision of dedicated task forces, UpP staff or students in agreement with universities), which allow the extraction of principles of law 46 from the measures issued by the courts of merit. The third area in which AI can be used is cataloguing, which allows maximised measures to be archived in easily accessible databases: whereas until a few years ago the only database was the CED of the Court of Cassation 47, today many judicial offices have begun to create their own databases, which are updated on a regular and continuous basis. AI can also be used to support any case law research in the databases, creating algorithms that can quickly identify precedents relevant to the case in question, thus enabling the sharing of knowledge between magistrates located throughout the country, but also between magistrates in the same judicial office. It seems illogical to allow law firms to be assisted by sentence search tools and, at the same time, prevent judges from the same judicial office from using systems that are specifically designed for the exercise of their functions and to assist them in their work 48. Finally, in the medium to long term, given the rapid pace at which AI is advancing a database could be created that is accessible to the entire community, designed to discourage the most prudent citizens from bringing reckless civil lawsuits by allowing them to consult recent decisions issued by the court in their district on similar issues 49.
To date, the use of AI in Italian civil proceedings is still limited and mainly experimental. The tools introduced so far operate in a limited number of pilot judicial offices and have not yet been systematically rolled out. At this stage, technological innovation has mainly resulted in models of assistance to the justice system, designed as tools to support the work of judges, with limited impact on the decision-making aspects of the exercise of judicial power, which remains entirely the responsibility and autonomy of individual magistrates. This is also because Article 15(3) of the new law on AI, i.e. Law No 132 of 23 September 2025, provides that, until the full implementation of the AI Act, i.e. EU Regulation 2024/1689, the testing and use of AI systems in ordinary courts must be authorised by the Ministry of Justice after consultation with the Agency for Digital Italy (AgID) and the National Cybersecurity Agency (ACN). Although in recent weeks the Ministry itself seems to have opened up to the possible introduction of AI systems, at least in civil proceedings, preferring to provide its own system rather than authorise the use of others. In fact, it has been decided that, starting from 1 January 2026, the Office 365 package provided to magistrates will include a licence for Microsoft Copilot 365, a generative artificial intelligence system developed by Microsoft. Although this is currently an experiment for purely informational purposes, on a voluntary basis only, and “(…) without any impact on the decision-making aspects of the exercise of judicial power, which remains entirely the responsibility and autonomy of the individual magistrate, it falls within the scope of support for the magistrate’s activity” 50. According to As a consequence, the Ministry believes that AI cannot replace judges, but judges are granted the right to make use of this augmented intelligence 51. This is also because the use of AI and jurimetrics 52 now seems unstoppable. Through the use of machine learning and cognitive analysis algorithms, it is not only possible to support magistrates in their work, but also to identify ex post recurrences in decision-making processes, prevailing jurisprudential trends and average times for the conclusion of proceedings, offering a statistical representation of the behaviour of the courts. The effect is that it is possible to estimate the probability of success of a given legal action, assess the advisability of appealing and guide litigation strategies on the basis of empirical indicators, marking the transition from supportive justice to progressively predictive justice 53.
Clearly, the question arises as to whether the use of such tools should really be a cause for concern and, if so, whether it can actually be prevented.
Firstly, it should be noted that using an artificial intelligence system does not mean delegating the entire management of the process to AI. The effectiveness of the digital ecosystem must be constantly monitored, even when AI is used for the simplest operations within the process: for example, it is necessary to ensure that the minutes and standards of the measures remain constantly aligned with sudden legislative changes or variations in the practices of individual judicial offices (e.g. in the event of a magistrate’s holiday or maternity leave). Furthermore, the use of AI by civil judges does not mean delegating decisions to the system itself: the judge always remains the sole decision-maker. Each application remains a tool to aid calculation or writing, unable, at present, to replace the critical and evaluative judgement of the judge. We cannot entrust the decision to AI in the same way that Judge Brigliadoca in Rabelais delegated the decision to a roll of the dice. The decision of AI may not be the most “excellent, honest, useful and necessary” decision 54; and this is true even if the AI has been perfectly programmed.
Furthermore, although the use of artificial intelligence raises concerns for some, it appears destined to become an unavoidable factor in our lives and therefore also in the pursuit of efficiency in civil justice. A change of perspective is therefore needed: given that the introduction of AI into the process is now a structural phenomenon, the question is not whether to fear its use, but how to govern its integration. This is also because it would be counterproductive for magistrates to renounce the use of such technologies, considering that lawyers already use them in defining defence strategies 55. The main legal challenge will therefore lie not in limiting, or even demonising, the use of AI, but in putting in place adequate regulatory and operational mechanisms capable of ensuring the safe use of AI tools, imposing a prudent and adequately regulated approach in order to mitigate risks and potential distortive effects. In any case, where such adverse effects occur, they can never be attributed to the machine, but rather to those who designed and programmed it. ChatGPT 56 agrees with this last sentence: when asked about it on the world’s most famous chatbot, it replied, “Attributing the error to the machine would be incorrect because AI does not interpret, evaluate or motivate: at most, it influences the decision. But influencing is not the same as deciding”.
Alfieri, A. (2022). L’ufficio per il processo: una possibile rivoluzione culturale nell’organizzazione degli uffici giudiziari. Giusto processo civile, 1197.
Ansanelli, V. (2014). Art. 191 c.p.c. In M. Taruffo, Istruzione Probatoria, Commentario del codice di procedura civile, Zanichelli.
Barbieri, F. (2022, 14 June). Il “nuovo” ufficio per il processo: un modello organizzativo aperto all’intelligenza artificiale. Giustizia insieme.
Biavati, P. (2019). Elasticità e semplificazione: alcuni equivoci. Rivista trimestrale di diritto e procedura civile, 1163.
Biavati, P. (2022). La riforma del processo civile: motivazioni e limiti. Rivista trimestrale di diritto e procedura civile, 45.
Braccialini, R. (2020, 1 June). L’ufficio per il processo tra storia, illusioni, delusioni e prospettive. Questione giustizia.
Calamandrei, P. (2019). Giustizia e politica: sentenza e sentimento. Opere giuridiche, I, RomaTrE-Press, 646.
Calamo, M., De Luzi, F., Macrì, M., Mencattini, T., Mecella, M. (2023). CICERO: A GPT2-based writing assistant to investigate the effectiveness of specialised LLMs’ applications in e-Justice. Frontiers in artificial intelligence and applications 372, 3196-3203.
Capponi, B. (2022). Note sulla fase introduttiva del nuovo rito di ordinaria cognizione. Giustizia civile, 2.
Capponi, B. (2023). Sulla fase introduttiva nel nuovo rito di ordinaria cognizione. Giustizia civile, 264.
Carpi, F. (2000). Processo civile e telematica: Riflessioni di un profano. Rivista trimestrale di diritto e procedura civile, 54, 470-482.
Carratta, A. (2020). Decisione robotica e valori del processo. Rivista di diritto processuale, 491.
Cavallini, D. (2021). L’ufficio per il processo. Rivista trimestrale di diritto e procedura civile, 981 ff.
Civinini, M.G. (2021, 8 September). Il “nuovo ufficio per il processo” tra riforma della giustizia e PNRR. Che sia la volta buona. Questione Giustizia, 3.
Consiglio Superiore della Magistratura (2023, 26 October). Organizzazione innovazione e statistiche. Il processo civile telematico, applicativi, consolle del magistrato.
Consolo, C. (2025). Spiegazioni di diritto processuale civile. Il processo di primo grado e l’impugnazione delle sentenze, II, Giappichelli, 228-230.
Costantino, G. (2004). Note sulla struttura della decisione nei processi a cognizione piena: Un modello per la consolle del giudice. Tecnologia, organizzazione e giustizia, 191-210.
D’Alessandro, E. (2025). Intelligenza artificiale generativa nel processo del lavoro: come cambia il rapporto tra avvocati e giudice. Lavoro, diritti, Europa, 1, 4.
Dalla Bontà, S. (2024). La cura delle parole tra processo e metodi consensuali: Per una gestione responsabile del conflitto. Giustizia consensuale, 1, 281-325.
Dalla Bontà, S. (2024). Mediazione e processo: Note a margine di recenti pronunciamenti. Judicium.
De Luzi, F., Macrì, M., Mecella, M., Mencattini, T. (2023). Cicero: An AI-based writing assistant for legal users. Intelligent Information Systems (CAiSE 2023), 477.
De Santis, A. D., Poli, G. G. (2013). Il processo civile telematico alla prova dell’obbligatorietà: lo stato dell’arte agli inizi del 2013. Foro italiano, 138(5), 109-120.
Delle Donne (2023). Art. 183 c.p.c. La riforma Cartabia del processo civile. In R. Tiscini (Eds.), Commento al d.lgs. 10 ottobre 2022, n. 149. Pacini giuridica, 298.
Di Cerbo, V. (2017). Banche dati di giurisprudenza, nomofilachia e trasparenza dell’attività giurisdizionale: L’esperienza del CED della Corte di cassazione. Questione Giustizia, 3, 93-98.
Fabbrini, F. (2013). Il processo civile telematico tra interpretazione del vigente e future evoluzioni. Giusto processo civile, 9, 271-285.
Fabiani, E. (2021). Intelligenza artificiale e accertamento dei fatti nel processo civile. Il giusto processo civile, 49, 62.
Fichera, G., Escriva, E. (2021, 4 February). Le quattro fasi del processo civile al tempo della pandemia. Judicium.
Fratini, F. (2024). Intelligenza artificiale e consulenza tecnica nel processo civile. Actualidad Juridica Iberoamericana, (20), 1200.
Gabellini, E. (2019). La «comodità nel giudicare»: la decisione robotica. Rivista trimestrale di diritto e procedura civile, 1305.
Gabellini, E. (2022). Algoritmi decisionali e processo civile: limiti e prospettive. Rivista trimestrale di diritto e procedura civile, 59.
Gentile, L. S. (2025). PNRR: Ancora misure straordinarie per l’organizzazione giudiziaria e la giustizia civile. Foro italiano, 150 (10), 431.
Gentile, S. L. (2022). Assunzione, inquadramento e mansioni degli addetti all’ufficio per il processo. Foro italiano, 5(9), 229-238.
Ghirga, M. F. (2022). L’ufficio del processo: una sfida. Rivista diritto processuale, 177 ff.
Giabardo, C.V. (2023). Ancora su «il giudice e l’algoritmo». Riflessioni critiche su intelligenza artificiale e giustizia predittiva (occasionate da un contributo di Michele Taruffo). Revista Ítalo-Española de derecho procesal, 1 ff.
Gioia, G. (2024). Chiari e sintetici: come scrivere in maniera efficace gli atti processuali secondo gli esperti. Pacini Giuridica, 4 ff.
Gioia, G. (2025). Intelligenza artificiale e indipendenza della magistratura: Un equilibrio tra trasparenza e controllo. Rivista di diritto processuale, 4, 1338-1363.
Irti, N. (2017). Per un dialogo sulla calcolabilità giuridica. In A. Carleo, Calcolabilità giuridica. Il Mulino, 19.
Irti, N. (2018). Il tessitore di Goethe (per la decisione robotica). Rivista di diritto processuale, 1177.
Lucarelli, P. (2015). Mediazione su ordine del giudice a Firenze: Prassi, problemi e linee guida di un modello. Utet giuridica.
Lucarelli, P. (2019). Mediazione dei conflitti: Una scelta condivisa. UTET Giuridica.
Lupoi, M. A. (1970). Giuscibernertica, informatica giuridica. Problema per il giurista. Raccolta di saggi sulla giurisprudenza, 32.
Magro Servet, V. (2018). The application of artificial intelligence in the administration of justice. Diario La Ley, 9268, 3-4.
Martínez Gutiérrez, R. (2021). Intelligenza artificiale, algoritmi e automazione nella giustizia. Proposte per una effettiva implementazione. Il Processo, 3, 437.
Merone, A. (2021). Le prove digitali e l’uso dell’intelligenza artificiale per finalità istruttorie. Giusto processo civile, 920-921.
Ministry of Justice (2022, 19 January). Relazione al Parlamento della Ministra della giustizia Marta Cartabia sull’amministrazione della giustizia. Giustizia.it.
Nieva Fenoll, J. (with Comoglio, P.) (2019). Intelligenza artificiale e processo. Giappichelli. (Original work published 2018).
Pagni, I. (2021, 17 November). L’ufficio per il processo: l’occasione per una (ulteriore) osmosi virtuosa tra teoria e pratica, con uno sguardo alle riforme processuali in cantiere. Questione Giustizia.
Picardi, N. (with Martino, R., Panzarola, A., Picardi, L.) (2019). Manuale del processo civile. Giuffrè Francis Lefebvre. (Original work published 1934).
Poli, G. G. (2014). Profili teorico-pratici del deposito degli atti nel processo civile telematico. Foro italiano, 137(5), 137-149.
Rabelais, F. (1994). Le tiers livre des faicts et dicts héroïques du bon Pantagruel (Chap. 39). In M. Huchon and F. Moreau, Œuvres complètes. Éditions Gallimard.
Santagada, F. (2020). Intelligenza artificiale e processo civile. Judicium, 467.
Sciacca, M. (2024). Sistema giudiziario e obiettivi PNRR: Il progetto cruscotto gestionale del presidente di sezione civile. Foro italiano, 5(1), 29-38.
Taruffo, M. (2018). Note sul precedente giudiziale. In A. Carleo, Il vincolo giudiziale del passato, Il Mulino, 109.
Testi, M. (2023). La mediabilità delle liti: Un modello operativo innovativo per l’efficienza della giustizia. In P. Lucarelli (Eds.), Giustizia sostenibile: Sfide organizzative e tecnologiche per una nuova professionalità. Firenze University Press, 245-259.
* PhD Candidate in Civil Procedural Law, University of Tuscia, Via Santa Maria in Gradi 4, 01100 Viterbo, Italy, alessia.bastianelli@unitus.it, ORCID ID https://orcid.org/0009-0002-2614-526X.
1 Recently, on the urgent measures introduced by Decree Law No. 117 of 8 August 2025, adopted in view of the final deadline of the National Recovery and Resilience Plan (hereinafter NRRP) scheduled for 30 June 2026, see Gentile, L. S. (2025). PNRR: Ancora misure straordinarie per l’organizzazione giudiziaria e la giustizia civile. Foro italiano, 150 (10), 431.
2 In general, on the development of electronic proceedings, see in doctrine Carpi, F. (2000). Processo civile e telematica: Riflessioni di un profano. Rivista trimestrale di diritto e procedura civile, 54, 470-482; Costantino, G. (2004). Note sulla struttura della decisione nei processi a cognizione piena: Un modello per la consolle del giudice. Tecnologia, organizzazione e giustizia, 191-210; De Santis A. D., Poli, G. G. (2013). Il processo civile telematico alla prova dell’obbligatorietà: lo stato dell’arte agli inizi del 2013. Foro italiano, 138(5), 109-120; Fabbrini, F. (2013). Il processo civile telematico tra interpretazione del vigente e future evoluzioni. Giusto processo civile, 9, 271-285; Poli, G. G. (2014). Profili teorico-pratici del deposito degli atti nel processo civile telematico. Foro italiano, 137(5), 137-149.
3 Fichera, G. — Escriva, E. (4 February 2021). Le quattro fasi del processo civile al tempo della pandemia. Judicium. https://www.judicium.it/wp-content/uploads/2021/02/Fichera-Escriva.pdf
4 For a commentary on the Riforma Cartabia, ex multis, see Biavati, P. (2022). La riforma del processo civile: motivazioni e limiti. Rivista trimestrale di diritto e procedura civile, 45; Delle Donne, C. (2023). Art. 183 c.p.c. La riforma Cartabia del processo civile. In R. Tiscini (Eds.), Commento al d.lgs. 10 ottobre 2022, n. 149. Pacini giuridica, 298; Capponi, B. (2022). Note sulla fase introduttiva del nuovo rito di ordinaria cognizione. Giustizia civile, 2.
5 The Ufficio per il processo’(hereinafter referred to as UpP) was established by Article 50 of Decree Law No. 90 of 24 June 2014 (converted into Law No. 114 of 11 August 2014), which introduced Article 16-octies of Decree Law No. 18 of 18 October 2012, No. 179 (converted by Law No. 221 of 17 December 2012), which at the time was intended to “(…) ensure the reasonable duration of proceedings through the innovation of organisational models and ensuring a more efficient use of information and communication technologies”. For the early stages of the UpP, see Braccialini, R. (1 June 2020), L’ufficio per il processo tra storia, illusioni, delusioni e prospettive. Questione giustizia. However, the UpP was only fully implemented with Decree Law No. 80 of 9 June 2021, containing “Misure urgenti per il rafforzamento della capacità amministrativa delle pubbliche amministrazioni nell’ottica di una pronta realizzazione del Piano nazionale di ripresa e resilienza per l’efficienza della giustizia”. On strengthening, see Civinini, M.G., (8 September 2021). Il “nuovo ufficio per il processo” tra riforma della giustizia e PNRR. Che sia la volta buona. Questione Giustizia, 3; Pagni, I. (17 November 2021). L’ufficio per il processo: l’occasione per una (ulteriore) osmosi virtuosa tra teoria e pratica, con uno sguardo alle riforme processuali in cantiere. Questione Giustizia.
6 Ministry of Justice (19 January 2022). Relazione al Parlamento della Ministra della giustizia Marta Cartabia sull’amministrazione della giustizia. Giustizia.it. https://www.giustizia.it/giustizia/it/contentview.page?contentId=ART362932.
7 The role of the UpP officer has a hybrid nature (see Ministry of Justice Circular No. 228522 of 3 November 2021), performing judicial support activities but also administrative support to the registries. Thus Alfieri, A. (2022). L’ufficio per il processo: una possibile rivoluzione culturale nell’organizzazione degli uffici giudiziari. Giusto processo civile, 1197; Barbieri, F. (2022, 14 June). Il “nuovo” ufficio per il processo: un modello organizzativo aperto all’intelligenza artificiale. Giustizia insieme; Cavallini, D. (2021). L’ufficio per il processo. Rivista trimestrale di diritto e procedura civile, 981 ff.; Gentile, S. L. (2022). Assunzione, inquadramento e mansioni degli addetti all’ufficio per il processo. Foro italiano, 5(9), 229-238; Ghirga, M. F. (2022). L’ufficio del processo: una sfida. Rivista diritto processuale, 177 ff.
8 Disposition Time (DT) is an international statistical indicator developed by the CEPEJ (European Commission for the Efficiency of Justice) to measure the average expected duration of legal proceedings. Unlike the actual duration of an individual trial, DT represents the ratio between the number of cases pending at the end of a given period and the number of cases concluded during the same period, multiplied by the time unit (usually 365 days). It expresses the capacity of judicial offices to dispose of their workload: a high DT value is symptomatic of a structural problem in the system and a tendency for backlogs to accumulate.
9 In particular, from north to south: Macroarea 1 — Next Generation UPP, Macroarea 2 — UNI 4 Justice, Macroarea 3 — Giustizia Agile, Macroarea 4 — MOD-UPP, Macroarea 5 — START UPP, Macroarea 6 — JUST-SMART.
10 On the use of AI in the process, see, among others, Irti, N. (2018). Il tessitore di Goethe (per la decisione robotica). Rivista di diritto processuale, 1177; Gabellini, E. (2019). La «comodità nel giudicare»: la decisione robotica. Rivista trimestrale di diritto e procedura civile, 1305; Carratta, A. (2020). Decisione robotica e valori del processo. Rivista di diritto processuale, 491; Santagada, F. (2020). Intelligenza artificiale e processo civile. Judicium, 467; Giabardo, C.V. (2023). Ancora su «il giudice e l’algoritmo». Riflessioni critiche su intelligenza artificiale e giustizia predittiva (occasionate da un contributo di Michele Taruffo). Revista Ítalo-Española de derecho procesal, 1 ff.
11 The project was coordinated by the University of Tuscia as lead partner, in collaboration with the universities of Rome Sapienza, Tor Vergata, Roma Tre, Cassino, Perugia, Florence, Pisa, Siena, the IMT School of Lucca and the Sant’Anna School of Pisa. The activity also involved 24 judicial offices belonging to the districts of Rome, Florence and Perugia. All reports prepared by the universities involved in the project are available at the following link: https://www.giustizia.it/giustizia/page/it/upp_dettaglio_macroarea_3
12 Sciacca, M. (2024). Sistema giudiziario e obiettivi PNRR: Il progetto cruscotto gestionale del presidente di sezione civile. Foro italiano, 5(1), 29-38.
13 The trial involved organisational analysts selected by FormezPA as part of the project Institutional Capacity Building of the Judicial Districts of the Sicilian Region, promoted by the Sicilian Region as part of the POR Sicilia FSE 2014-2020 Axis IV—OT11—Specific Objective 11.4.
14 The report on the Giustizia Agile project, drawn up by the University of Florence, can be consulted at the following link: https://www.giustizia.it/cmsresources/cms/documents/3giusagile_unifi_modorgrepupp_gagile_dinfo.pdf
15 The Court of Cassation, on the subject of separation between spouses, has ruled that the right to receive maintenance allowance pursuant to Article 156 of the Italian Civil Code is based on the persistence of the duty of material and moral assistance. It is therefore related to the standard of living maintained during the marriage and, unlike the divorce allowance, does not have any compensatory components. On the contrary, the divorce allowance has a welfare and compensatory nature due to the definitive termination of the bond. On this point, see Civil Cassation, 7 January 2025, no. 234; Civil Cassation, 18 September 2025, no. 25556.
16 To understand how the mechanism for calculating the allowance developed during the Giustizia Agile project works, see the report published on the Ministry of Justice website, available at the following link: https://www.giustizia.it/cmsresources/cms/documents/3giusagile_unipg_supdig_sup_tecnologico_causa.pdf.
17 The use of the Milan Tables as a criterion for calculating compensation for non-pecuniary damage was affirmed by the Supreme Court in its well-known ruling Civil Court, Section III, 6 May 2020, no. 8532.
18 The Consolle del Magistrato is a Java application and is the only programme that allows judges and their assistants to manage the Processo Civile Telematico (PCT): through this programme, judges can view all their cases and the information and data for each procedure in the court registry, as well as draft orders using MS Windows Word, sign them digitally and send them to the registrar for publication in the electronic file. On this point, see Consiglio Superiore della Magistratura (2023, 26 October). Organizzazione innovazione e statistiche. Il processo civile telematico, applicativi, consolle del magistrato. https://www.csm.it/web/csm-internet/il-processo-civile-telematico/consolle-del-magistrato.
19 The parenting plan is an example of an organisational model for the civil sector drawn up by the University of Tuscia and submitted to the Court of Viterbo as part of the Giustizia Agile project. To consult the model, see the following link: https://www.giustizia.it/cmsresources/cms/documents/3giusagile_unitus_modorgciv_piano_genitoriale.pdf
20 This mapping was made possible through the use of the Inspectors Package, an application that includes a set of software developed by the Directorate-General for Automated Information Systems (D.G.S.I.A.) and which allowed the data contained in the civil sector’s computerised registers to be extracted in real time.
21 For the new structure of civil proceedings, see Capponi, B. (2023). Sulla fase introduttiva nel nuovo rito di ordinaria cognizione. Giustizia civile, 264; Consolo, C. (2025). Spiegazioni di diritto processuale civile. Il processo di primo grado e l’impugnazione delle sentenze, II, Giappichelli, 228-230.
22 This refers to the software developed as part of the ‘Next Generation’ project of Macroarea 1—North—West Italy and funded by PON Governance e Capacità Istituzionale 2014-2020.
23 Testi, M. (2023). La mediabilità delle liti: Un modello operativo innovativo per l’efficienza della giustizia. In P. Lucarelli (Eds.), Giustizia sostenibile: Sfide organizzative e tecnologiche per una nuova professionalità. Firenze University Press, 245-259.
24 Dalla Bontà, S. (2024). La cura delle parole tra processo e metodi consensuali: Per una gestione responsabile del conflitto. Giustizia consensuale, 1, 281-325.
25 Lucarelli, P. (2015). Mediazione su ordine del giudice a Firenze: Prassi, problemi e linee guida di un modello. Utet giuridica; Lucarelli, P. (2019). Mediazione dei conflitti: Una scelta condivisa. UTET Giuridica; Dalla Bontà, S. (2024). Mediazione e processo: Note a margine di recenti pronunciamenti. Judicium.
26 Testi, M. (2023). La mediabilità delle liti: Un modello operativo innovativo per l’efficienza della giustizia. In P. Lucarelli (Eds.), Giustizia sostenibile: Sfide organizzative e tecnologiche per una nuova professionalità, Firenze University Press, 245-259.
27 More specifically, various projects have been carried out by the University of Florence in collaboration with the Umbrian and Tuscan judicial offices since 201: the ‘Nausicaa’ project and the ‘Giustizia Semplice’ project in the Court of Florence, the project in the Court, the ‘Dike’ project in the Court of Appeal of Florence, the ‘Giustizia Condivisa’ project in the Court of Perugia and the ‘Jacobea’ project (2020-2021) in the Court of Pistoia.
28 For an analysis of the relationship between AI and fair trial, see Gabellini, E. (2022), Algoritmi decisionali e processo civile: limiti e prospettive. Rivista trimestrale di diritto e procedura civile, 59.
29 Gioia, G. (2025). Intelligenza artificiale e indipendenza della magistratura: Un equilibrio tra trasparenza e controllo. Rivista di diritto processuale, 4, 1338-1363.
30 Fabiani, E. (2021). Intelligenza artificiale e accertamento dei fatti nel processo civile. Il giusto processo civile, 49.
31 Lupoi, M. A. (1970), Giuscibernertica, informatica giuridica. Problema per il giurista. Raccolta di saggi sulla giurisprudenza, 32.
32 On the general doctrine of court-appointed technical consultants in civil proceedings, see Ansanelli, V. (2014). Art. 191 c.p.c. In M. Taruffo Istruzione Probatoria, Commentario del codice di procedura civile, Zanichelli, 9. Sull’utilizzo di sistemi di IA nella nomina del consulente tecnico v. Nieva Fenoll J. (with Comoglio P.) (2019). Intelligenza artificiale e processo. Giappichelli, 66. (Original work published 2018); Fabiani, E. (2021), Intelligenza artificiale e accertamento dei fatti nel processo civile. Il giusto processo civile, 62; Fratini, F. (2024). Intelligenza artificiale e consulenza tecnica nel processo civile. Actualidad Juridica Iberoamericana, (20), 1200.
33 Nieva Fenoll, J. (with Comoglio P.) (2019). Intelligenza artificiale e processo. Giappichelli, 66. (Original work published 2018).
34 Nieva Fenoll, J. (with Comoglio P.) (2019). Intelligenza artificiale e processo. Giappichelli, 66. (Original work published 2018).
35 Fabiani, E. (2021). Intelligenza artificiale e accertamento dei fatti nel processo civile. Il giusto processo civile, 49.
36 To consult the report prepared by the University of Catania as part of the Project, please visit the following link: https://www.giustizia.it/giustizia/page/it/upp_dettaglio_macroarea_6
37 Art. 35-bis of Legislative Decree 25 of 28 January 2008.
38 This is stated by Merone (2021). Le prove digitali e l’uso dell’intelligenza artificiale per finalità istruttorie. Giusto processo civile, 920, interpretando N. Picardi (with Martino, R. Panzarola, A., Picardi L. (2019). Manuale del processo civile. Giuffrè Francis Lefebvre, 128. (Original work published 1934).
39 Thus A. Merone (2021). Le prove digitali e l’uso dell’intelligenza artificiale per finalità istruttorie. Giusto processo civile, 920-921.
40 As proposed in France by Information Report No. 216 (2024-2025) on AI and the legal professions.
41 On the potential of using AI in the decision-making phase, see most recently Gioia, G. (2025). Intelligenza artificiale e indipendenza della magistratura: Un equilibrio tra trasparenza e controllo. Rivista di diritto processuale, 4, 1338-1363.
42 On the requirement for conciseness in civil proceedings imposed by the Riforma Cartabia, see G. Gioia, G. (2024). Chiari e sintetici: come scrivere in maniera efficace gli atti processuali secondo gli esperti. Pacini Giuridica, 4 ff.
43 Calamo, M., De Luzi, F., Macrì, M., Mencattini, T., Mecella, M. (2023). CICERO: A GPT2-based writing assistant to investigate the effectiveness of specialised LLMs’ applications in e-Justice. Frontiers in artificial intelligence and applications 372, 3196-3203.
44 For the report prepared by the University in the context of the Giustizia Agile project, please refer to the following link: https://www.giustizia.it/giustizia/page/it/upp_dettaglio_macroarea_3
45 De Luzi, F., Macrì, M., Mecella, M., Mencattini, T. (2023). Cicero: An AI-based writing assistant for legal users. Intelligent Information Systems (CAiSE 2023), 477.
46 Taruffo, M. (2018). Note sul precedente giudiziale. In A. Carleo, Il vincolo giudiziale del passato, Il Mulino, 109; Irti, N. (2017). Per un dialogo sulla calcolabilità giuridica. In A. Carleo, Calcolabilità giuridica. Il Mulino, 19.
47 On the contribution of legal informatics to the nomophilic function of the Court of Cassation, see Di Cerbo, V. (2017). Banche dati di giurisprudenza, nomofilachia e trasparenza dell’attività giurisdizionale: L’esperienza del CED della Corte di cassazione. Questione Giustizia, 3, 93-98.
48 Martínez Gutiérrez, R. (2021). Intelligenza artificiale, algoritmi e automazione nella giustizia. Proposte per una effettiva implementazione. Il Processo, 3, 437.
49 On the importance of means of ascertaining the courts’ positions before the start of proceedings, especially in labour matters D’Alessandro, E. (2025), Intelligenza artificiale generativa nel processo del lavoro: come cambia il rapporto tra avvocati e giudice. Lavoro, diritti, Europa, 1, 4.
50 To consult the circular of 9 January 2025 that launched the Copilot system in judicial offices, see the link: https://pst.giustizia.it/PST/resources/cms/documents/Nota_informativa_sperimentazione_Copilot_signed.pdf
51 Fabiani, E. (2021). Intelligenza artificiale e accertamento dei fatti nel processo civile. Il giusto processo civile, 49.
52 With regard to legal metrology, see the work of Magro Servet, V. (2018). The application of artificial intelligence in the administration of justice. Diario La Ley, 9268, 3-4.
53 Shouldn’t decisions always be predictable? Calamandrei, P. (2019). Giustizia e politica: sentenza e sentimento. Opere giuridiche, I, RomaTrE-Press, 646. On predictability, see Biavati, P. (2019), Elasticità e semplificazione: alcuni equivoci. Rivista trimestrale di diritto e procedura civile, 1163.
54 Rabelais, F. (1994). Le tiers livre des faicts et dicts héroïques du bon Pantagruel (Chap. 39). In M. Huchon and and F. Moreau, Œuvres complètes. Éditions Gallimard.
55 However, following events in the US where some lawyers cited non-existent judgments, they are now obliged to acknowledge that they used AI in drafting their documents. See the US case v. Merken S. (9 June 2023). Another US judge says lawyers must disclose AI use. Reuters.com, https://www.reuters.com/legal/transactional/another-us-judge-says-lawyers-must-disclose-ai-use-2023-06-08/
56 ChatGPT is an artificial intelligence system developed by OpenAI, based on GPT (Generative Pre-trained Transformer) language models.