European Parliament Artificial Intelligence | The European Parliament gives the green light to the project to regulate artificial intelligence

2023-06-14 15:57:25

The first comprehensive law on Artificial intelligence of the world is a little closer. He European Parliament This Wednesday has become the first assembly in the world to establish the contours of the new regulation, a pioneer worldwide, the content of which will begin to negotiate with the Council and the European Commission this Wednesday. “There is no time to lose”, said the industry commissioner, Thierry Breton. The European Parliament wants the remote biometric surveillance on public roads in real time is totally prohibited, that the generative AI systems such as ChatGPT have to notify that content has been generated by AI and that AI systems used to influence voters in elections are considered high risk.

“We send a clear message to the world of what needs to be done, how we can preserve democracies while preserving innovation”, highlighted the report’s co-sponsor, the Italian Social Democrat Brandon Benifei. The text has received majority support when it was approved with 499 votes in favour, 28 against and 93 abstentions. The proposal launched by the European Comission in April 2021, it establishes a risk-based approach, with obligations for providers and those who deploy AI systems based on the level of risk that this technology may generate.

For this reason, he proposed to prohibit AI systems with an unacceptable level of risk for the safety of people, such as those used to classify people based on their social behavior or personal characteristics. The European Parliament now demands that this prohibition be extended to other intrusive and discriminatory uses of AI such as remote biometric identification systems ‘in real time and ‘after the event’ in public access spaces. The idea has gone ahead despite the last-minute attempt by the European People’s Party (EPP) to allow the use of mass surveillance by the police in cases of disappearance of minors and terrorism. It will only be possible after the fact and with judicial authorization.

Parliament also proposes to ban the biometric categorization systems that use sensitive characteristics (for example, gender, race, ethnicity, citizenship status, religion, or political orientation), predictive policing systems (based on profiles, location, or past criminal behavior), emotion recognition in police services, border management, the workplace and educational institutions, as well as the non-selective extraction of facial images from the Internet or video surveillance recordings to create facial recognition databases because they understand that it violates human rights and the right to privacy.

“AI raises many questions from a social, ethical and economic point of view. But now is not the time to hit any ‘pause button’. On the contrary, it is a question of acting quickly and assuming responsibilities”, Breton has valued, who has stressed that with the vote this Wednesday the European Parliament becomes the first parliament in the world to demand an “exhaustive regulation” of Artificial Intelligence. Since the new rules would not enter into force until 2026, the EU wants to negotiate a grand AI deal with the US in the meantime, as announced by Vice President Valdis Dombrovkis. Breton, as he has announced, will travel to San Francisco in the next few days to meet with the creator of Meta, Mark Zuckerbergy Sam Altman de Open AI.

Inter-institutional negotiations

The approval of the European Parliament is not, in any case, the end of the road, although it does open the door at the beginning of interinstitutional negotiations with the Council. The first inter-institutional three-party dialogue will take place this Wednesday in Strasbourg, although it will be the Spanish presidency of the UE from July 1 who assumes the leadership of negotiating this “important file” with the aim of closing an agreement before the end of the year. “I hope the key points will include clear and proportionate rules on generative AI. We need effective transparency requirements on content generated by AI and strict rules against ‘Deep fakes’”, Breton claimed.

The regulation also proposes that systems that can be used in different applications are analyzed and classified according to the risk they pose to users. In this way, the different levels of danger will imply more or less regulation. In the classification of high-risk applications are AI systems that pose significant harm to people’s health, safety, fundamental rights or the environment. Also those systems used to influence voters and the outcome of the elections as well as the recommendation systems used by social media platforms (with more than 45 million users).

The regulations requested by the European Parliament also require providers of foundational models, an innovation in the field of AI, to have the obligation to evaluate and mitigate possible risks (for health, safety, fundamental rights, the environment, democracy and the rule of law) and register their models in the EU database before placing them on the market in the EU.

As for generative AI systems based on those models, such as the ChatGPT, they will have to comply with additional transparency requirements that include the identification of the content as generated by AI. The goal is to make it easier to distinguish fakes from authentic images. They will also have to publish detailed summaries of the copyrighted data that has been used in their development. MEPs also want citizens to be able to file complaints about these systems and to guarantee their right to obtain explanations about decisions generated by high-risk systems that significantly undermine fundamental rights.