"AI systems are tools to help legal professionals, but they cannot become more important than people"

Technology companies facing the challenge of managing DATA and AI' was the title of the conference given by Miguel Ángel Acero, director of Digital Transformation Strategy and Technological Innovation at Izertis at DATAFórum Justicia, the sector's annual forum, which took place in Granada in 2024. In this interview, he talks about the challenges of Justice, the role of AI and Generative AI in this field and how Izertis is helping in these processes of change.

1.How is Izertis helping in the digital transformation of justice and public administration in general towards the world of data management?
-As a global technology consultancy, we are working in several areas. On the one hand, we are helping public institutions design their strategy and adopt data platforms that facilitate the organisation, storage and analysis of large amounts of information. In this way, we achieve improved accessibility to this data and also centralised and secure management, deploying data governance strategies in an agile way. We are working on the implementation of data management platforms, both structured and unstructured, that integrate Generative AI capabilities that facilitate the interaction and consumption of stored information in an intuitive way.

Izertis also supports institutions in complying with data protection regulations such as GDPR and other specific regulations, working to ensure that citizens' confidential data is handled ethically and securely. Alongside this, there are other issues associated with cyber-security systems and procedures in this area. 
Another new focus of work and service is the deployment of predictive solutions or the implementation of citizen service systems based on conversational platforms such as RASA, where Izertis is the leading partner worldwide. 

2.Linked to the deployment of digital transformation solutions, what challenges remain to be met in Justice and what can be done to achieve them?
-We are facing challenges with technological, organisational, social and regulatory aspects. I would like to focus on the former. The Administration of Justice manages a huge amount of unstructured information, processing tens of thousands of proceedings annually, often consisting of dozens of pleadings. We are talking about tens of millions annually. Accessing all this information in an accessible and simple way, both for professionals and citizens, is for me one of the main challenges. This requires a variety of initiatives to promote this access. For example, improving interoperability between systems, establishing common standards for all justice systems and platforms, based on open protocols, or promoting the use of APIs, which allow the seamless integration of systems and platforms, ensuring that information flows without barriers between the different parts of the justice system.

Another important aspect is data security and protection, through the implementation of advanced cybersecurity protocols or creating a robust data governance framework to ensure the integrity, privacy and confidentiality of judicial information.

It is essential to develop more intuitive and accessible platforms for the public

3.Cybersecurity and improving interoperability. Another challenge is the simplification of procedures? 
-It is very important to work on accessibility and simplification of accessibility and simplification of judicial procedures for citizens. Although electronic platforms such as the Electronic Court File (EJE) exist, many citizens still find it difficult to interact with the judicial system digitally, especially those with fewer technological resources or unfamiliarity with digital tools. This is why I believe it is imperative to develop platforms that are more intuitive and accessible to the public, with user-friendly interfaces and simplified procedures. 

For example, a platform such as RASA is essential to implement more agile and accessible citizen services that help citizens carry out procedures remotely. At the same time, it enhances e-inclusion by offering support to users who need help accessing e-Justice services, through a channel that is very natural to them such as Whatsapp.

4.It is also essential that the algorithms used in the administration of justice are transparent and explainable, isn't it?
-That's right. Citizens have a right to understand how decisions that affect their lives are made, especially when it comes to judicial processes. At Izertis we are working on the development and adoption of Artificial Intelligence (AI), with the implementation of AI systems that are transparent and explainable. As stated in the policy on the use of AI, measures must be implemented in the Administration of Justice to ensure that automated decisions are understandable and can be examined by experts and stakeholders. I do consider that we can experiment with state-of-the-art AI services which, for reasons of novelty, are not capable of offering high explainability, provided that they do not undermine the rights of litigants, or the judicial functions themselves, and are fully compliant with the applicable regulatory regime.

Finally, I believe that the provision of a GovTech Justice Strategy that serves to consolidate a culture of collaborative innovation (between the judicial and technological sectors) is also key, promoting greater participation of the innovation and entrepreneurship ecosystem, both in the identification of challenges and in the development of solutions, supported by emerging technologies.

5.How can Artificial Intelligence help the Administration of Justice meet the challenges it faces and what do you think could be an optimal way to address them?
-We are all aware that the judicial system must make the most of technology and specifically of Artificial Intelligence, but always attached to ethics and bearing in mind that the administrator of justice is ultimately a human being. It is essential that AI systems are used as tools to assist legal professionals in their functions, but the ultimate responsibility for making legal decisions must rest with judges and magistrates. 

Their independent decision-making must be ensured. The human review of everything generated whenever it directly or indirectly affects the rights of the users of the public service of justice or the jurisdictional activity itself. We cannot allow the tool to become more important than the person.
We must create 'human-in-the-loop' systems, with human oversight in AI decision-making processes to be included in any algorithmic platform. Even if an artificial intelligence system is used, it must be ensured that judicial decisions can be reviewed beforehand. 

Users believe that ChatGPT is highly reliable when it is not

6.Review and monitoring procedures must therefore be included
-Human intervention can override decisions that the AI system does not 'understand' will be detrimental to users. The level of human supervision required, as well as the timing of supervision, depends on the level of risk involved and the potential implications of delay.
Another aspect for access to justice informed by AI is the transparency and explainability of algorithms. Without a clear understanding of the factors involved in the systems, the basis for a decision or the ability to challenge it after the fact, due process is at risk.
We must also inform citizens who may be affected by the system, in clear and understandable language whether the results provided by the tool are binding or not, as well as their possible use during the judicial procedure, before or during it, and their right to object so that they can be heard directly by a judicial body.

The use of AI systems also increases the capacity to develop long-term sustainable, inclusive and reliable policies and services. In addition, it provides relevant information to public justice managers to enable them to make decisions on planning, development and evaluation of appropriate legislative policies, as well as on organisational and structural planning issues.

7.What is your perspective on the impact that Generative Artificial Intelligence can have on the field of Justice? What is Izertis developing along these lines?
-In the latest LexisNexis Legal & Professional® Report worldwide practitioners convey that they see the greatest potential for generative AI tools as helping with researching matters (65%), drafting documents (56%), analysing documents (44%) and drafting communications and emails (35%). But also consumers and citizens are largely using generative AI to research legal issues (60%), obtain information to develop a will (40%), create a company (37%) or develop a rental contract (39%).

This is a cocktail that if not managed correctly can have a very strong impact on our lives. Users believe that ChatGPT has high reliability when in fact it does not.. Educating users about the capabilities and limitations of AI models, as well as providing clear guidelines on how to use them effectively and responsibly, can help mitigate risks. Generative AI should be seen as an element that helps us to complete automation processes incorporating a differential value for the global deployment of solutions. 

8.Can you give us an example of its use?
-We can start, for example, from one of the IA services implemented by the Ministry of Justice, such as the anonymisation of documents. The use of this tool has the function of anonymising the information in documents that can identify a person for the purpose of tracing him or her, or using his or her data fraudulently. This is a service that can be developed with NLP techniques prior to AI Gen and that can also be enriched with a document classification service with NLU-based autocataloguing systems, allowing significant time savings and simplifying the tasks of the professionals involved. However, by incorporating AI Gen solutions, it is possible to further develop the service of automated summarisation of court documents in plain language. This service aims to improve the understanding of judicial processes, guaranteeing the accessibility and proximity of the Public Justice Service. Finally, the incorporation of an intelligent search engine allows legal professionals to search for the information they require in legal documents using simple language. 

This example, clearly marks the impact that AI Gen can have on judicial management systems. On this last point of being able to facilitate the clear language of judicial communications, I believe it has an almost immediate impact, where technology already allows us these capabilities with the integration of specific RAGs and ensuring that people can easily understand the content and implications of judicial decisions.

GEN AI is the AI tool that will contribute intensively to the realisation of citizens' right to clear and comprehensible information. Bias in AI results often stems from biased training data. It is crucial to ensure that training data sets reflect diversity in the many dimensions that are important for access to justice (e.g., race, ethnicity, income, education). Without a wide range of demographics, perspectives and scenarios, any AI tool is unlikely to serve its intended user base.