What to expect (and what not) from the regulation of Artificial Intelligence?

It is often said that the first thing to do when you are going to do something is to lower your expectations;

because, if they are set too high, it will only generate disappointment.

And something like that happens with the European Artificial Intelligence regulation and, in general, with any regulation.

Because there are many people who still believe that laws solve problems and those of us who dedicate ourselves to it know that they do not.

But, at the same time, we know that a regulatory framework is necessary.

First of all, we must congratulate that a political agreement has been reached and we must congratulate the Spanish presidency and, especially, the Secretariat of State for Digitalization and Artificial Intelligence (SEDIA) for the effort made in a negotiating marathon of 36 hours, from December 6 to 8.

It was important that the European Union was the first region in the world to have a general standard on the use of AI.

And it has been achieved.

But, despite the importance of the agreement, the processing of the regulations has not been completed.

It is now in a technical negotiation phase and will have to be definitively approved by the European Parliament and the Council.

Afterwards, it will have to be translated into all EU languages, so we can talk about several months.

And, once published, there will be a period of two years for its application, although the application of some parts is planned to be brought forward (creation of the European AI Office, list of prohibited uses, etc.)

On the other hand, the European AI regulation is not and does not aspire to be a complete and exhaustive regulation of this, but, rather, a 'regulatory framework' that allows development through delegated acts of the European Commission, guidelines of the European AI Office that is created and the AI ​​Supervision Agencies of the Member States and also the Good Practice Agreements that are signed with the sectors, to be able to adapt to the innovations that are produced in this field.

And there will also be modifications in the sectoral substantive laws.

For example, although agreement has been reached on basic rules that should govern the use of generative AI, solutions to the many doubts and intellectual property problems raised by this technology will have to be addressed within the framework of Property regulations. Intellectual.

directive in the Digital Single Market

will be reviewed soon , surely the movements and pressures to make legislative changes will begin at the national level (in Spain they have already started).

Regarding 'biometric surveillance' and other dangerous uses of AI for the Fundamental Rights of citizens, the important thing is not only the list of prohibitions and exceptions, but, above all, establishing who is going to monitor those who they watch

A system of guarantees for citizens and sanctions must be provided for authorities and public administrations that misuse AI (including the disqualification of senior officials and officials, apart from other responsibilities they may incur), so that what does not happen The same thing happened with the General Regulation and the Spanish Data Protection Law, in which it was decided not to fine Public Administrations.

In any case, we must always keep in mind that the purpose of the rules is to regulate conduct to avoid risks and conflicts, imposing obligations and sanctioning non-compliance with them.

And, in the case of the AI ​​Regulation, it is about trying to prevent and avoid or, if this is not achieved, to sanction the misuses that are made of it, which is why it can convey a negative idea of ​​it, when, perhaps , we should talk more about the good uses and their benefits.

For this reason, both the European Artificial Intelligence Regulation, which is a first general legal framework, and the sectoral standards that are developed from it, must be approached with a positive and flexible attitude, so that the regulation does not impede innovation and , above all, so that bad uses (which, of course, must be monitored, prevented and, where appropriate, punished), do not make us miss out on their good uses.

Borja Adsuara Varela

is an expert in Law and Digital Strategy.