Article in El País: “Artificial intelligence agents: an untimely loophole in European regulation”

On June 25, 2025, the financial newspaper Cinco Días (El País) published an article by Pablo García Mexía, Director of Digital Law at Herbert Smith Freehills Kramer, entitled "Artificial intelligence agents: an untimely loophole for European regulation."

In it, Pablo analyzes the emerging phenomenon of artificial intelligence agents: systems built on generative models capable of planning, creating code, and acting in the real world autonomously, without direct human intervention. This enhanced autonomy multiplies both their potential and their social risks.

The article emphasizes that, while the challenges of privacy, discrimination, and intellectual property are not new, the big difference is that these risks now materialize through decisions made autonomously by the machine on behalf of individuals. This raises a dizzying scenario in terms of legal liability.

Although the European Artificial Intelligence Regulation (AIR) approved in 2024 is applicable to these systems, a major regulatory gap arises because the European Union decided not to address civil liability arising from the use of AI in that regulation, leaving this issue to parallel legislation.

Following the withdrawal of the draft Directive on AI civil liability in 2025, damage caused by AI agents will be subject to the civil regimes of each Member State, making it extremely difficult for victims to obtain compensation.

The conclusion is clear: AI agents, although not representing a new technological paradigm, will be the next major social upheaval in artificial intelligence. And without a clear European liability framework, this innovation threatens to create serious tensions both for the protection of citizens and for the legal certainty of businesses.

Read the full article in Cinco Días: https://cincodias.elpais.com/legal/2025-06-25/agentes-de-inteligencia-artificial-una-inoportuna-laguna-para-la-regulacion-europea.html