Miguel Ángel Presno Linera

On October 22, I had the opportunity to participate in the Second Artificial Intelligence Week organized by the Faculty of Law at the University of Seville. I shared a session with Professors Stefano Pietropaoli and Roger Campione, moderated by Professor David Sánchez Rubio, where we discussed the ethical-legal challenges as well as the opportunities presented by artificial intelligence (hereinafter, AI). I am grateful for the generosity of Professor Fernando Llano, Dean and the driving force of this event, which reached its second edition in 2024.

In my presentation, I referred to ten «challenges» and five «opportunities.” Regarding the challenges, I briefly commented on the following:

The first challenge is to dismantle the recurring objection that the regulation of AI is, in any case, a constraint on research, development and innovation in this field, and that it is best to let the companies developing these systems to determine where the limits should be set. The AI that already exists presents real risks, some of extraordinary magnitude, which cannot be ignored by public authorities and must be addressed, either by subjecting these systems to thorough oversight or, when the risk is unacceptable, by prohibiting them altogether.

The second challenge to overcome is the notion that a purely ethical approach is sufficient to address the dangers of AI. Legal intervention is essential and it is imperative to achieve a comprehensive and adequate regulatory framework.

The third challenge is to promote an interdisciplinary legal approach, so that AI regulation considers the various legal fields affected by these systems, including constitutional, criminal, administrative, tax, European, public and private international, civil, commercial, labour, procedural and legal philosophy.

But, fourthly, we cannot limit ourselves to an exclusively legal approach; rather, cooperative work is needed with mathematics, computer science, medicine, physics, chemistry, economics, philology, sociology…

The fifth challenge is to regulate systems that are extraordinarily dynamic, requiring a combination of precise rules regarding what can or cannot be done, along with a set of principles, goal-oriented norms that prescribe the objectives public authorities must achieve while allowing them the flexibility to determine the means to do so.

Sixth, it is not easy to delineate what will be regulated and what will remain outside the framework or subject to future regulation. Two recent European regulations (the European Union Regulation and the Council of Europe Framework Convention) have defined the systems they apply to as «a machine-based system designed to operate with varying levels of autonomy and capable of demonstrating adaptability after deployment. For explicit or implicit objectives, it infers from the input information it receives how to generate output results, such as predictions, content, recommendations or decisions, which can influence physical or virtual environments». Thus, algorithmic systems that lack this autonomy, such as VioGén, fall outside the regulatory scope of these norms.

The seventh challenge to address is the risk that AI systems may exacerbate existing digital, social, cultural and economic divides, in other words, elitist AI.

Eighth, it is essential to avoid the extraordinary cost that the development of AI systems may impose on many fundamental rights and constitutionally protected assets. In addition to the already well-known misuse of private data and images to train these systems, other examples include their impact on the right to literary, artistic, scientific and technical creation, as well as their significant consumption of energy and water, which affects natural resources and the environment.

The ninth challenge is to ensure that public authorities do not use these systems to reinforce their own immunity.

Finally, and in tenth place, there is a significant risk that public bodies established to oversee and, where necessary, sanction violations committed by those who develop and use these systems may be captured and «neutralized» by the very entities they are meant to regulate, particularly by major AI corporations.

Regarding the opportunities that AI offers, the first one is the streamlining of administrative procedures that can be automated without creating any vulnerabilities for the individuals affected by their processing.

The second opportunity is to promote an interdisciplinary approach to problems, based on the necessary interdisciplinarity that AI demands.

Third, it is worth considering that the caution that must be taken to prevent biased AI system design may serve to highlight the fact that prejudices based on race, gender, age, social or economic status… are present in our societies and must be removed not only in the digital environment.

Fourth, the legal regulation of AI should contribute to the restoration, at least in this domain, of the primacy of the law in contrast to the growing expansion of private contracts, as highlighted by Professor Esteve Pardo in a recent book. These contracts have progressively taken over areas that traditionally and constitutionally belong to that source of law.

Fifth, since AI is already here and is becoming increasingly relevant, we have the opportunity to establish a genuine «right to artificial intelligence». This would be understood as a complex right, not explicitly mentioned in the constitutional text but derived from various constitutional provisions — including the mandate of Article 18.4 of the Constitution, which states that the law shall limit the use of information technology to guarantee the honour, personal and family privacy of citizens, and the full exercise of their rights. This right would encompass a variety of safeguards in the form of individual freedoms, prohibitions on public authority and entitlements. Among them: the right not to suffer discrimination by AI systems and to access the benefits derived from them; the protection of physical and moral integrity, as well as ideological freedom, against potential intrusions posed by these systems; the guarantee of respect for rights linked to private and family life, self-image, and to literary, artistic, scientific and technical creation; the protection of freedom of movement and freedom of expression against the misuse of surveillance systems based on facial recognition; the preservation of effective judicial protection and the right to a defence in the context of an expanding «judicial artificial intelligence;» Digital literacy; the guarantee of decent working conditions and the right to unionization and strike action in the face of the increasing algorithmic power of corporations; the protection of consumer rights…

*This translation has been revised by María Amparo González Rúa from the original Spanish version, which can be consulted here.