In its future regulation on artificial intelligence (AI), the European Commission wants to ban the use of intrusive algorithms that aim to establish widespread surveillance or manipulate human behavior . However, this exclusion, as we have recently learned, will not apply to public bodies when security reasons are alleged.
The European Commission presented its proposal for a regulation on artificial intelligence on April 21, as a continuation of its white paper, in which it presented its strategy on AI, in order to establish an “ecosystem of excellence” and an “ecosystem of trust ”to promote artificial intelligence, especially taking into account some of its associated risks, also including increased investments or algorithmic bias about which we have already spoken to you on other occasions.
At that time, the first version of the white paper was officially published on February 19, 2020, where the Commission advocated an approach based on both regulation and investment , with a double objective: to promote the use of artificial intelligence, on the one hand, and taking into account the different risks associated with certain uses of this new technology, on the other.
Precisely, for the creation of an “ecosystem of excellence”, the Commission has requested that cooperation between the different European research centers be strengthened, in addition to proposing the creation of a center to coordinate all efforts, attracting investment and researchers “more talented in the discipline ”.
The use of AI for widespread surveillance would be banned
According to different sources, it seems that Brussels wants to ban AI applications that are considered too intrusive . This would mean that it would prohibit the systems used to carry out an “indiscriminate surveillance applied in a generalized way to all natural persons without differentiation”.
And the same would happen with those algorithms designed to manipulate human behavior, or to rate people based on their actions.
Should any company violate this principle, the companies would risk a fine of up to 4 percent of their annual turnover , in accordance with the proposed European regulation.
However, public authorities and governments of the European Union shall have the right to repeal this rule for the preservation of public safety. Although no definition has been given for this, it would have as its main objective activities related to terrorism or organized crime.
The importance of regulating what the European Commission considers high-risk AI
As can be seen from the text presented, the European Commission wishes to strictly regulate what it calls “high risk” artificial intelligence , as could be the case of facial recognition technologies installed in spaces easily accessible to the public, establishing new requirements for evaluation.
In addition, it also provides for the implementation of a European Artificial Intelligence Council , made up of a representative from each of the 27 Member States, a European Data Protection Supervisor and a representative from the European Commission itself.
The supervisor, for example, would be responsible for issuing relevant recommendations and opinions to the Commission, regarding the list of artificial intelligence practices.
However, for the moment we will have to wait for the arrival of the different proposed measures. Not surprisingly, once the executive has presented them, both the Council and the European Parliament itself must approve them.
Although some MEPs have already advanced that they consider that the rules are not strict enough. And some of them have already asked the Commission to address the fundamental rights risks posed by so-called “high risk” AI. Ursula von der Leyen, current President of the Commission, has promised that this request will be considered.
