The European Parliament has adopted this morning the negotiating position on the Artificial Intelligence Regulation. But the position adopted by the MEPs places Parliament and the Commission in opposite camps regarding the processing of biometric data in real time.
The Commission considers that there should be more guarantees for criminal investigation, with the possibility of processing in real time for certain data in special circumstances. Parliament is more in favor of guaranteeing privacy.
The rapporteur for the first law on artificial intelligence, MEP Brando Benifei, assumes that the parliamentary debate was not without controversy and defends that legislation must go hand in hand with technology.
“Today [quarta-feira] we had a last-minute disagreement in Parliament on a point about maintaining a clear ban on real-time biometric identification,” said the deputy, regretting “the attempt to politicize the issue, [e] to turn it into a propaganda tool.”
The president of the European Parliament, Roberta Metsola, defends that MEPs must ensure that technological developments respect European values: “As legislators, we have to seize the opportunity, it is about change, understanding that we cannot afford to remain stagnant and not have fear of the future.”
“In the future we will need clear and constant borders and limits for artificial intelligence”, warned the president of the European Parliament, speaking with the rapporteurs of the approved text, Brando Benifei, who wanted to ensure that these limits will be defended by Parliament.
“We managed in Parliament to maintain a clear safeguard to avoid any risk of mass surveillance,” said Benifei, who guarantees that “at the same time, the possibility will be maintained, with biometric identification in non-real time, of pursuing criminals and any risks that we have in society”.
For her part, the Competition Commissioner, Margrethe Vestager, considers that there should be exceptions and, in the first reaction to the vote in Parliament, defended that some biometric data should be able to be processed in real time.
“For example, if the police are looking for a missing child or if there is a terrorist situation and there are people on flights who need to be found,” Vestager said, acknowledging that “Parliament does not agree with this position.”
“I think that the Council is currently very much in line with the Commission’s proposal,” he stressed, a few hours before the negotiations (trilogues), which begin this Wednesday night. “In the negotiations we will see how it is resolved,” Vestager stressed, with the expectation that by the end of the year the European Union will approve the first law on artificial intelligence.
Risk and prohibited for AI
“The standards follow a risk approach and establish obligations for both providers and users, depending on the level of risk that AI can generate,” says the European Parliament in a note released after the vote.
“AI systems with an unacceptable level of risk to the safety of people will be strictly prohibited, including systems that implement subliminal or deliberate manipulation techniques, exploit people’s vulnerabilities, or are used for social scoring (ranking people). people according to their social behavior, socioeconomic status, personal characteristics)”, can be read in the same note.
Among the list of practices that Parliament wants to ensure that they will also be prohibited is “the recognition of emotions and predictive surveillance”.
“Generative AI systems, such as ChatGPT, must make it clear that the content was generated by artificial intelligence,” defend the MEPs, who also consider that the AI data “used to influence voters in elections is high risk” .
Source: TSF
