What Can We Expect from the New EU Artificial Intelligence Law?

The European Union’s Artificial Intelligence Act is a milestone. However, it is worth asking what impact it will have and whether it meets the expectations it has generated. We will analyze it in this article.

The European Union has been the most agile institution when it comes to legislating the development of artificial intelligence. The sudden irruption of this technology into our lives made it necessary to accelerate the processing of legislation on this matter, in order to safeguard the rights of citizens and avoid the risks that could arise from its uncontrolled use.

After several years of work, and coinciding with the Spanish Presidency of the EU, the European Parliament and the Council reached an agreement last December on the basic points of the new Artificial Intelligence Law. Although the main aspects of the new regulation had been defined for months, as we reported on Silicon.es.

The text that has finally been announced establishes four levels of risk around the use of artificial intelligence. And it also specifies the fines in case of non-compliance, which we already reported.

Positive assessment

Experts are satisfied with the agreement reached. “The law is a step forward for the protection of human rights against invasive or unethical practices in Europe, while seeking to promote technological development. At a time of expansion of artificial intelligence, it lays the foundations to avoid possible malpractices that may affect individual freedom, such as biometric surveillance or influence on voting,” says Sonia Martínez Requejo, professor of Education and Educational Innovation at the European University.

Likewise, José Antonio Pinilla, president and CEO of Asseco Spain Group, stresses that “the main focus is on implementing protection measures and restrictions for the use of biometric identification systems in public environments”.

He also remarks that “it is optimistically expected that this new legislation will play a crucial role in balancing technology with civil rights, thus seeking to greatly mitigate any negative impact on society”.

Martinez Requejo says the law meets the expectations that the industry had. “We have been able to access previous drafts and regulatory approaches for some time and there have been no surprises. In my opinion, the regulation is quite balanced and coherent with previous texts on artificial intelligence in European territory”, he assures.

In any case, we will have to be attentive to its development. “This draft regulation must be translated into legal language, where requirements, measures and sanctions will be specified. Given its pioneering nature, it represents a novelty for companies operating on the continent. Overcoming the initial phase of expectation will be the main challenge in adapting economic activity to the new needs of European users,” Pinilla explains.

He also recalls that “this is a new regulation, of which we have no precedents, so I am sure that, as technology advances, new regulations will be necessary”. Thus, he emphasizes that “public institutions will have to closely follow the progress in technological innovation to be able to legislate accordingly”.

“The important thing at the moment is the predisposition to regulate, since, in this first step where changes are constant, it would be unfeasible to be able to mark a regulation that anticipates what may happen in the coming months,” he acknowledges.

A first step

The European University professor stresses that “it is a pioneering law, which will influence others and which chooses to put people at the center of technological innovation”. “Best practice recommendations aimed at companies developing artificial intelligence models are not enough to guarantee the transparency of these models or the fundamental rights of citizens. A common regulatory framework is needed to establish the mechanisms to ensure that artificial intelligence is at the service of people and not the other way around,” he specifies.

On the other hand, CEO Asseco believes that “the EU’s initiative in introducing this law not only has implications at the regional level, but also sets a global precedent.” “This law stands as a model to follow for regulation in countries outside the European community, which will surely be watching and taking note of the measures that will be put in place. The intention is that these non-EU countries can imitate and adapt the key elements of this law, in order to avoid being left behind in comparison with other markets,” he says.

In this way, he believes that “the influence of this European regulation is not only limited to its borders, but is projected as a benchmark in the creation of international policies and standards related to the implementation of technologies such as artificial intelligence”.

In short, he considers that “the EU not only seeks to regulate its own space, but also contributes to guide and promote ethical and responsible practices at a global level“.

How will it affect companies?

This law sets certain limits on the use of artificial intelligence, so companies will have to adapt. “It establishes a clear framework for the development and use of artificial intelligence. This helps companies establish clear processes that help reduce limitations such as biases or security and/or data management issues. It also establishes a framework for cooperation between companies, researchers and regulators. This can help stakeholders work together to develop and implement artificial intelligence in a safe and ethical manner,” notes Martínez Requejo.

“In the near future, companies will have to adapt to the new paradigms of the sector and will have to comply with the new obligations established, which will vary depending on the use and relationship with artificial intelligence,” Pinilla points out. Although he admits that “it is still too early to estimate the consequences that this new law could have on companies that already operate with artificial intelligence models”.

For the moment, it should be noted that, in our country, in November the Council of Ministers approved a Royal Decree empowering the Ministry of Economic Affairs and Digital Transformation, through the Secretary of State for Digitalization and Artificial Intelligence, to open the call for companies to participate in the controlled testing environment -sandbox- promoted by the European Regulation, as the expert from the European University recalls.

Gradual implementation

It should be noted that the new European law will not fully come into force until 2026, as they will be implemented in phases.The European office will be set up immediately. The ban on high-risk models will come in six months. And the need to meet requirements for generative artificial intelligence systems and models, 12 months. In addition, the Commission will publish a report evaluating and reviewing the proposed artificial intelligence framework five years after the date of entry into effect of the framework, to ensure that it is up to date and responsive to societal and technological demands,” specifies Martínez Requejo.

Two years is perhaps too long for a technology that is evolving as fast as artificial intelligence. “It is still too early to know whether the law will have to be reformed before it comes into effect in 2026. However, it is common for laws and regulations, especially those related to emerging technologies such as artificial intelligence, to be subject to adjustments and revisions as they are implemented and practical experience is gained,” concludes Asseco’s CEO.