A project to regulate Artificial Intelligence (AI) in the EU passed a key stage on Thursday get the first green light from MEPs, who called for new bans and for the ChatGPT phenomenon to be addressed in greater depth.
MEPs from Parliament’s Civil Liberties and Consumer Protection committees voted overwhelmingly in favor of limit the use of AI in the European Union (EU), while continuing to encourage innovation in the sector.
(Also: Europe sets its eyes on regulating ChatGPT with the artificial intelligence law)
The text will will present to the plenary session in June for its adoption and if approved, negotiations will begin with the Member States in order to agree on a final law.
The European Commission had already proposed a AI regulation law two years ago; an ambitious regulatory project in which member countries defined their positions at the end of 2022.
But the parliamentary review has dragged on and been delayed in recent months as controversy erupted over the dangers of AI tools capable of creating text or images.
(Also: Teacher, programmer and other careers at risk from artificial intelligence)
MEPs described Thursday’s vote as “historic” and stated that they hope it will end with the implementation of “the world’s first regulation on artificial intelligence”.
The approved text includes some of the main axes of the proposal made two years ago by the European Commission, but also suggests a ban For Biometric Surveillance, emotion recognition and AI systems for predictive policing.
Handling risks
“We received more than 3,000 amendments. You just have to turn on the television, every day we see how important this issue is for citizens“, said the Romanian Liberal MEP Dragos Tudorache, one of the promoters of the text.
Another of the promoters of the draft, the Italian social democrat parliamentarian Brando Benifei affirmed that “Europe wants an ethical approach, based on the human being”.
(Read: ChatGPT: Could artificial intelligence take away your job?)
Europe wants a
ethical approach, based
in the human being
AI systems are highly technically complex and These tools generate fascination, but also concerns.
On one side, can save lives by allowing an important advance in medical diagnoses, but also can be used by authoritarian regimes to implement mass surveillance of the population.
Public opinion discovered its immense potential late last year with the launch of the ChatGPT text generator robot, from the Californian company OpenAI, which is capable of writing essays, poems or translations in a few seconds.
But the spread on social networks of fake images that look very realistic created with applications like Midjourney generated alert about the risks of manipulation of public opinion.
(We recommend: ChatGPT AI is the biggest technological advance in decades, according to Bill Gates)
Amid heated debate, a group of leading scientists called for a moratorium on the development of more powerful systems until there is better legislative regulation.
The text approved this Thursday is based on the current regulations on product safety and it is going to impose controls that fall mainly under the responsibility of the companies.
(You can read: Joe Biden requires technology companies to make artificial intelligence safe)
Human control over the machine
The core of the draft is to create a list of rules that are imposed only on applications considered as “high risk”a qualification that must be declared by the companies themselves, based on the criteria established by the legislators.
For the Brussels executive this includes all systems used in areas such as critical infrastructure, education, human resources, order control or migration management.
(It may interest you: Tech giants will discuss AI ‘risks’ at the White House)
The best way for the EU to inspire other jurisdictions is to ensure that new regulations enable, not inhibit, the development of useful AI applications.
One of the obligations sought to be imposed is that there be a control by a human of the machine, the establishment of technical documentation and the creation of a risk management system.
Compliance with these rules would be the responsibility of the supervisory authorities designated by each State of the bloc.
(More news: The hard blow that the Supreme Court of Mexico gave to Amlo’s electoral reform)
The European Parliament is also seeking further consideration of content-generating AI tools such as ChatGPT, calling for the establishment of a specific regime of obligations of the same type than the one applicable to high-risk systems.
The CCIA, a lobby group defending the big digital companies, reacted by stating that while “Parliament made some improvements to the text”, it was, on the other hand, “abandoning the risk-based approach” of the Commission proposal. European.
“The best way for the EU to inspire other jurisdictions is to ensure that the new regulation enables, and does not inhibit, the development of useful AI applications”said Boniface de Champris, the CCIA’s European policy director.
AFP
#European #Union #free #rein #law #regulate #Artificial #Intelligence