The future of Artificial Intelligence in the European Union

A critical analysis of the Proposal for a Regulation on Artificial Intelligence and its coexistence and compatibility with the Proposal for a European Regulation on the protection of personal data. SCIENTIFIC PREMISES: In the fourth Industrial Revolution, Artificial Intelligence is an important and fundamental technology, and occupies a not inconsiderable place in the future development of humanity, its recognition is based on the need to facilitate scientific progress, achieve technological leadership, ensure that new technologies are at the service of all citizens, but at the same time these respect the established fundamental rights and improve the quality of life of these, by improving it and also contributes to business development and services of public interest. The impact of Artificial Intelligence systems must always be seen in two perspectives, one individual perspective and another as a society as a whole, they are inextricably linked. Its regulation from the legislative point of view is fundamental to provide a safe, reliable development of Artificial Intelligence, which respects the values and fundamental rights of citizens, with an ethical use of them and whose fundamental pillars are excellence and trust, not only of artificial intelligence providers but also of their users. In both cases they must always be seen not from the perspective of each of the member states of the union, but as a union itself, being also a driving force to achieve the sustainable development goals foreseen in the 2030 agenda and the objectives foreseen in the Green Pact. It is visible the need to regulate, from the legislative point of view, artificial intelligence, because it is inextricably linked to data. There is no artificial intelligence without data. What makes it necessary to coexist and make compatible with the established regulations on data protection with the established regulations on

In this sense, we believe that those relating to electoral processes should be included as high-risk AI systems, nor does it apply to technologies in the field of advertising, nor that it would affect consumer privacy in general terms.
In this sense, a situation of urgency duly justified in the third paragraph of said article is also regulated, which in no way is defined, but to which, in addition, is added the counterproductive situation that after the AI system is used, the authorization can be granted, when already at that time such a concession would not make any sense, so in our opinion, although it is not wrong for such a situation to occur, the truth is that for this to happen, in no way can the concession be established after its use.I also consider in this regard the inclusion of the prohibition of the use of AI systems that cause or are expected to cause damage to the environment in any of its forms.
In the specific case of the requirements established in Chapter 2 for high-risk AI systems, it is proposed in relation to paragraph 8, that this be a prohibition and not a suggestion, if we take into account the repercussions that this may have on minors and always taking into account that this would guarantee compliance with the postulates of the Convention on the Rights of the Child.
Article 15 whose purpose is to make clear the conditions of precision, robustness, and cybersecurity that must meet high-risk AI systems, in its Section 4 must regulate other conditions that may arise in terms of vulnerability and what to do in these situations.
The White Paper suggests that in the regulatory framework of intelligence systems as part of the Trust Ecosystem, within its requirements that of record and data conservation, this was established in the Regulation of Artificial Intelligence specifically for high-risk AI systems in Article 12 in relation to 20, which in our opinion should be expanded and deepened taking into account the complexity and opacity of many of these systems and the difficulties that may arise in this regard, which would allow effective verification of compliance with the applicable rules and execute them, because in said Article in the first order, as we had already mentioned, this requirement is established for high-risk AI systems, without mentioning the rest of the AI systems, such as those with low or minimum risk, which due to certain failures in their systems can suddenly become high risk, in a second order it is wise to regulate the period of time that these records will be kept, which must be limited and reasonable, but must be defined, all in order to facilitate the monitoring and verification of the actions and decisions of these systems.Moreover, the conservation of documentation on programming methodologies, process training, and techniques used to build, test, and validate AI systems should be regulated.Here we could also add the conservation and recording of the data set used to train and test AI systems.
In the specific case of human surveillance established in Article 14 of the regulation, or what is the same human supervision, in addition to the objectives set out in its paragraph (2) it must be added at our discretion that this will also be in order to prevent or minimize risks to the environment, making this particular very clear because of the impact it has on the very life of the human being an individual entity and as a social entity.In addition, it would modify in paragraph (5) of that Article the number of natural persons established to confirm and verify an odd number and in the specific cases of paragraph (4) in subparagraphs (d) and (e) it should be made clear that the intervention can occur in real time, effectively ensuring that the AI system does not undermine human autonomy or cause adverse effects.
Article 20 specifies that providers of high-risk AI systems shall keep log files for an appropriate period of time, and this lawyer's criterion must be clearly defined, what that period of time would be In the case of Article 56 concerning the European Artificial Intelligence Committee, a third paragraph would be added concerning the supervisory function of the same on the national supervisory authorities in terms of the correct application of the regulation, as it contains these harmonised rules for the whole of the European Union.
The legislation lacks regulation in terms of neurorights, that is, in terms of privacy and mental identity, which has a direct impact on the decision-making of the individual, as well as no issues related to social networks or fake news, all of which are linked to information.
In the technical field, the systems known as privacy-preserving machine learning are also not regulated.
Special mention should be made in this legislation regarding the use of AI in intellectual creations, as well as the civil liability is not regulated SECOND: There is harmonization of the regulations on Artificial Intelligence and the regulation on the protection of personal data, they complement each other, even many of the concepts defined in both as biometric data, personal data, and others correspond in their definition in both instruments, but indisputably the second is much more comprehensive in terms of effective judicial protection, because it establishes resources, responsibility and sanctions before the control authority when the first is only limited to establishing sanctions consisting of administrative fines, but does not establish the resources that can be established against them.In this sense it is clear that the RGPDP comprises a set of concepts in which those referring to personal data, generic data, health-related data and in the regulation on AI the definitions of terms such as those mentioned above are ratified, but it goes further in relation to defining concepts such as training data, validation data, test data, input data, which lead to the real need for a regulation on data in a general way and does not specify as the RGPD, which has as its center the personal data.Of particular significance in this is the regulation of Article 54 of the AI regulation, which includes the further processing of personal data for the development of certain AI systems for the public interest in the controlled testing space for AI, being evident that only in this sense such processing is regulated, being disregarded in relation to the treatment at other times in the life of AI systems that turn out to be of special significance.There is even talk of completely novel terms such as non-personal data, anonymised data, synthetic data.
There is extreme concern as to how to address data privacy and the relationship with AI, the GDPRP is designed to protect the privacy of all citizens of the European Union and give them control over their personal data, in it it was made very clear that they must coexist with transparency and with privacy.AI lives on huge amounts of data, its processing falls within the scope of the GDPRP, so every time an AI system uses personal data, consequently all the provisions of that r egulation apply to this system.It is very clear that in the phases of creation and use of dand AI systems personal data are used, this being a vital element for the complete life cycle of AI systems, BUT not all AI systems use personal data, there are systems that usedata that are not personal, being very difficult to define the line between personal and non-personal data.
Article 5 of the aforementioned RGPDP establishes the principles relating to the processing of personal data, summarized in the number of 6, all transcendental for AI systems and in this regard has special importance the fact that it must be finalized from the beginning before the entry and operation of AI systems, the purpose of the use of the aforementioned personal data and can not be used differently, as well as will take into account the limitations established to the processing of data by the legislation on the protection of personal data in its Article 23 and following, they must be treated in a tender manner, loyal, transparent, for specific, explicit, and legitimate purposes, they must be accurate and up-to-date, treated in such a way as to ensure adequate security and maintained in such a way as to allow the identification of data subjects for no longer than necessary, which transcends the AI systems that use them.To do this, we must: know the regulations that relate to both AI systems and those derived from them, do not use means that deceive or carry out fraudulent actions for the processing of personal data, being respected the interests of the owner, as well as the privacy that the owner of the personal data will always want, as well as the full knowledge that he must have of what will happen to them.As for the purpose it should be made clear that AI systems can use personal data and that this will be only for the initial purpose conceived, to create techniques for reviewing this data so that they are always accurate, complete, and up to date.
The processing of personal data using AI systems must be based on the lawfulness of the processing established in Article 6 of the RGPDP, based on the 6 conditions established by it, and must be regulated both in the training phaseand in the use phase, that is, in its processing phase and in its interaction with customers, administrators and other systems or applications Consequently, the fact that Article 9 of the GDPR establishes a treatment of special categories of personal data is related to the real and certain fact that AI systems are used in all sectors of human life, so it would be governed by the exceptions established in Section 2 of the aforementioned Article itself.
Our analysis and proposal would be aimed at reconciling in AI systems the treatment that is given to personal data in the different stages of life of said systems, as well as the principles that govern this, the legality of said treatment, and its compatility with all the treatment given to AI systems with that given to the protection of personal data always under the view that not all the data used in the aforementioned systems are of that nature and on the basis of the emergence of other types of data that currently lack regulation and that it is up to this law to define and prosecute them in such a way as to improve the legality of AI systems.
It is necessary to prepare a system or guide that reconciles the legal system in terms of the treatment of personal and non-personal data with AI systems, there would be our research, combining it with the impact on decision-making and its link with the management of our emotions and feelings, in order to make possible the development of AI with the right to the protection of personal data.
We believe we are at the right time to ensure that AI technologies comply with the rules set by the RPGDP to ensure people's privacy and the ability to self-regulate the information they wish to share.We must seek a balance between technological advances and the need for the protection of personal data, then in turn we will be protecting fundamental rights and values.That is possible.


Journal the Law.January 22, 2020.United States presents 10 principles for the development of Artificial Intelligence  Official Journal of the European Union.May 4, 2016.RegulationEU 2016/679 of the European Parliament and of the Council of 27 April 2016 on the protection of individuals with regard to the processing of personal data and the circulation of such data  Douglas Heaven Will.The big week against biased AI: THE EU and THE US begin to regulate  Jimenez Marimar.Published in the newspaper El País-The EU establishes red lines to AI why are they and who they affect? García José.14 de Enero del 2020.China affirms that a text written by an AI is protected by copyright. Gudi Faustino.August 2020.The Libror White of the European Commission the attempt to achieve technological humanism  Artificial Intelligence Law dated April 21, 2021. White Paper on Artificial Intelligence, a European approach aimed at excellence and trust  OECD.Paris May 22, 2019.Forty-two countries adopt the OECD Principles on Artificial Intelligence  Parra Sergio.March 2, 2020.La European Union publishes guidelines to regulate Artificial Intelligence  Drafting Types.July 19, 2021.Lights and shadows of the new European law  Reuters.July 13, 2021.The United States will seek global rules to prevent the misuse of Artificial Intelligence  Corner Andreu Gerard.European Commission White Paper on Artificial Intelligence.A European approach to excellence and trust. Sánchez Alonso, Aida.April 21, 2021.Four levels of risk to AI in the proposed European legislation  Sánchez del Campo Alejandro.Garrigues Digital.This is how the United States wants to regulate artificial intelligence  Digital site Andina Link-January 13, 2021-The change of position of the United States on Artificial Intelligence. The Technolawgist.May 10, 2021.AI regulation: current proposals in the US, EU and UK  Valdecantos Martane.May 13 of the 2021.La artificial intelligence, at bay in the EU  Vega Guillermo.17january 2020.EU plans to ban facial recognition in public places for up to 5 years Msc.Alena Pérez Urdanivia Lawyer