A starting point so far unique in the world and a great challenge: this is the AI Act approved by the European Parliament according to artificial intelligence experts, jurists and philosophers.
There are many bets on the table: from the necessary harmonization of the way in which European countries will accept the regulation, to the ability to keep pace with the speed with which artificial intelligence advances and changes, up to the inevitable social consequences in the sectors of employment and of the new computer literacy that appears necessary from now on.
Rotolo (Icsc), innovative and ambitious approach, a common strategy is now needed
A starting point and an innovative and ambitious approach, but a common strategy on artificial intelligence is now needed: this is how he commented on the approval of the AI Act by the European Parliament Antonino Rotolo coordinator of the transversal research group of the ICSC national center and deputy director of the Alma Human AI Center of the University of Bologna.
“The AI Act is a very complex regulation, which is intended to be comprehensive, regulating all aspects of artificial intelligence, from development to supply and use,” he observed.
From this point of view it is a "very innovative" step.
The underlying theme, he added, "is risk".
That is, it is about "ensuring that problems relating to artificial intelligence are reduced to a minimum, thanks to a security control".
One of the big bets is keeping up with innovation.
From this perspective, Rotolo observes, the AI Act was concerned with finding a definition of artificial intelligence in line with a flexible regulatory system.
In fact, it is a question of keeping pace not only with rapid technological development, but also with the development of governance at a global level.
"In this sense, the Ai Act is useful for positioning Europe in a process of global harmonization": it is a "very ambitious step and "an important stoic step".
It is however a first step because some points will have to be explored in depth For example, although the regulation is "well written, very clear and rigorous, some passages could generate different interpretations within the member states", observed Rotolo. The second point concerns the governance and interaction of the national authorities with those EU and the third point concerns the verification and compliance processes in the development of artificial intelligence systems.
Bertolini (Eura), the challenge is security, important pieces to implement in the coming years
"An epochal document" which is the first in the world to take up the challenge of the safety of products linked to artificial intelligence and which in the future will have to be implemented and constantly updated: this is what Andrea Bertolini, director of the AIAct, comments on the AiAct approved today by the European Parliament. European excellence on the regulation of robotics and artificial intelligence, Eura, and professor of private law at the Scuola Superiore Sant'Anna.
The Ai Act "is a regulation" and, as such, after its entry into force it is applicable in all EU countries.
This should guarantee greater uniformity, even if the applications in individual countries will be different", he observed. "It is the first regulation in the world that deals with artificial intelligence" and, as happened for privacy, "it pursues the effect Brussels, that is, the EU tries to establish rules: anyone who wants to bring innovations to Europe will have to adapt to these rules". It
would be a mistake, Bertolini said, "to think that regulating means prohibiting: the aim is to encourage good innovation". It is above all a "discipline on product safety", a sort of certification of products in relation to safety" on the basis of "criteria and standards which partly exist and which will have to be developed. Today - he noted - the document has been approved in the final formulation, but over time standards will have to be adopted on the criteria to comply with and emerging technologies will have to be taken into consideration. Important pieces will have to be implemented in the months and years to come."
Always looking to the future, according to the expert, new scenarios can be imagined.
For example, a "market for the certification of products based on artificial intelligence" could arise and "probably even those who innovate will have to get used to the idea of consulting jurists already in the planning and design phases to understand how to adapt to regulatory requirements".
De Caro (Roma Tre), a step that must not remain isolated
"A very important step because the problems that increasingly accelerated technological progress brings with it cannot be addressed at the level of individual nations": thus the philosopher Mario De Caro, of Roma Tre University comments on the approval of the AI Act by the European Parliament.
These are "global problems", which would require "a collective awareness of all nations: proceeding in no particular order would lead to less satisfactory results".
The possibilities offered by artificial intelligence, such as facial recognition or generative models for producing tests and images "offer enormous opportunities, but also pose enormous risks".
In this sense, the "European standards are encouraging".
Their objective, she notes, is "to protect our security against forms of illegality that were unknown until a few years ago and were not regulated".
The important thing is that the European AI Act "does not remain an isolated case: it is a first step and more will need to be done on several fronts, with global work", especially considering that "there is almost daily progress and that the work must be continually updated."
Looking to the future, according to De Caro the challenge does not only concern the limits for technologies: "we need to think about how to reorganize society to deal with any distortions", he added thinking of the "impressive quantity of jobs that will disappear and be replaced in the coming years by artificial intelligence. We will have to ask ourselves how to re-employ people and furthermore, for the first time, technological progress threatens intellectual jobs. This - concluded the philosopher - does not mean that there will not be new jobs, but there will be people without work for long periods. Education must enable people of all ages to be trained in new technologies. It is the new challenge."
Reproduction reserved © Copyright ANSA