You only have a minute? Here are 3 takeaways from this article:
- The European Parliament has voted in favor of the European AI law, marking an important step towards establishing a European regulation on AI.
- The regulation commits to banning AI systems that pose an unacceptable level of risk to personal safety, such as AI-powered facial recognition.
- While researchers praise the proposed law, startups are expressing concerns about its enforcement and the actual benefits it would bring to the industry.
The vote was adopted with 499 votes in favor, 28 against and 93 abstentions, with the next steps now being negotiations on the final form of the law. According to For the co-rapporteurs of the proposals, the EU is making history with this vote and has opened the way to a dialogue which will ultimately also involve the rest of the world.
“We made history today – we paved the way for the dialogue we need to begin with the rest of the world about how we can build responsible AI for our planet and the systemic risks this can bring. But I also think of ordinary citizens, consumers, businesses and institutions who need to be supported in the adoption of AI so that they can get the most out of it. In addition, they can be sure that they can trust that the institutions have put in place a system of safeguards capable of identifying real risks,” said co-rapporteur Brando Benifei, of the Progressive Alliance of Socialists group. and Democrats.
Prohibition of intrusive and discriminatory use of AI
Facial recognition proved to be the most contentious issue in the vote, with MPs debating the extent to which biometric surveillance should be limited. MEPs also rejected a proposal from the EPP group that the risks posed by real-time biometrics in public spaces could be offset during “extraordinary circumstances”.
“We won in Parliament to maintain a clear guarantee to avoid any risk of mass surveillance, and at the same time to maintain the possibility, without real-time biometric identification, of pursuing criminals and all the risks that we run in society . » Benifei added.
As a result, the prohibited list will now be include intrusive and discriminatory uses of AI such as:
- “Real-time” remote biometric identification systems in publicly accessible spaces;
- “Post” remote biometric identification systems, with the sole exception of law enforcement for the prosecution of serious offenses and only after judicial authorization;
- Biometric categorization systems using sensitive characteristics (e.g., gender, race, ethnicity, citizenship status, religion, political orientation);
- Predictive policing systems (based on profiling, location or past criminal behavior);
- Emotion recognition systems in law enforcement, border management, workplace and educational institutions;
- Untargeted scraping of facial images from the Internet or CCTV footage to create facial recognition databases (violating human rights and the right to privacy).
For co-rapporteur Dragos Tudorache of the Renew Europe group in the EP, one of the most important aspects of the text provides clear protections against such controversial use of AI.
“The text provides protections for how seriously we have taken bans on the use of artificial intelligence, how seriously we have considered high-risk applications, and mechanisms to also ensure that we are not unnecessarily limiting how high-risk applications could be used. be forced to comply,” Tudorache said.
THE rule book also specifies that model-based generative AI systems such as OpenAI’s ChatGPT or Google’s Bard should comply with transparency requirements. It should be disclosed that the content was generated by AI, which would also allow so-called “deep-fake” images to be distinguished from real images and provide safeguards against the generation of illegal content.
“We also serve the agenda of promoting innovation, without hindering creativity, as well as the deployment and development of AI in Europe, which is an objective just as important as that of protecting our citizens,” he said. declared the Romanian MEP.
A positive step in the right direction or a growing concern?
According to the researchers, the EU AI law is becoming the first truly global effort to regulate AI, and a positive signal that focuses first and foremost on safety and security issues.
“It is quite incredible that we have not yet understood the need for such a framework. Another positive point is that it not only tackles AI made in the EU, but also outsourced parts of AI systems, which is a huge thing. It could be said that this even goes beyond the borders of the EU. This need is of course justified by the protection of EU fundamental rights and by safety and security within the Union,” said an AI researcher based in Italy. Viktor Milochevsky tells The Recursive.
However, there are still challenges ahead, particularly regarding its application, adds Miloshevski.
“The keywords surrounding the entire AI Act experience are prohibited artificial intelligence practices, classification of AI systems as high risk, standards, conformity assessment, certificates , registration and governance. The keyword “app” is also a pretty important keyword. And this is where things could get a little tricky. At this stage of design at Union level, the proposal establishes a European Artificial Intelligence Committee (the “Council”), composed of representatives of Member States and the Commission. In my opinion, enforcement would be quite difficult in the first two years, with an increasing tendency towards further adaptation,” says the AI researcher.
For Marko PorobijaManaging director of Croatia-based law firm Porobija & Spoljaric, the proposed regulation prioritizes security and ethical considerations, which is essential to ensure that “the benefits of AI are shared fairly and that the consequences potential negative effects are minimized.
“I appreciate the focus on transparency and accountability, as this will help build trust in AI systems and ensure they are developed and used responsibly. Overall, I think the EU Artificial Intelligence Act is a positive development that will promote the responsible use of AI and protect people’s well-being,” Porobija told The Recursive.
However, some are also concerned about this law. According to David MengerCEO of a conversational AI startup based in Prague Wingbot.aiIn its current form, the AI Act is “a patchwork law.”
“The only thing that makes sense is banning facial recognition in public places. The law completely ignores really serious issues, such as the use of AI for fraud, counterfeits and scams, as well as privacy and copyright in terms of training data. Many things, like fraud laws and data copyright, are barely enforceable. This is just another legal hurdle for businesses, not a solution to real problems. This could really mess up AI startups in the EU,” Menger told The Recursive.
Depending on the final form of the law, Menger also worries about what the regulation would do for end users of different AI products and services.
“I’m concerned that the time and money we would like to spend on creating better products is being spent on complying with the law, and our users are not seeing any real benefit,” he adds.
For Raluca Apostol, co-founder and chief product officer of human intelligence platform Nestor, when it comes to privacy and security risks, there was guidance in this area even before the rise of data tools. Generative AI.
“In Europe, under the GDPR law, compliance policies and secure processing of personal data are stipulated very clearly, with provisions in force before 2018. Although the GDPR encompasses some aspects relevant to generative AI, such as data protection, consent and transparency, there are additional considerations specific to generative AI that may require additional regulations or guidelines. Yet there is a framework,” Apostol tells The Recursive.
However, there are many challenges and concerns that need to be addressed.
“These are real concerns that need to be further addressed through comprehensive regulations and standards. Unfortunately, we have to balance a double-edged sword here. On the one hand, as a business, you don’t want your sensitive data transferred to untrusted systems. On the other hand, blocking your employees’ access to AI tools can leave you behind your competitors. So I don’t think there is an easy answer to these challenges or a magic solution to solve these problems in the near future,” she emphasizes.
What’s next for AI law?
The European Parliament must now begin further negotiations on the final text in tripartite negotiations with the Commission and EU member countries.
According to Tudorache, this will be a process that will need to include everyone and will require a coordinated effort.
“In addition to writing rules, our governments and ourselves, here at European level, must engage in a very serious and coordinated effort to explain and support our citizens in this transformation. This means that we need to invest heavily in education and reskilling our current workforce to be able to integrate AI into their work, because otherwise we would have forces in our societies that would resist and watch this transformation of ‘an evil eye – and that’s it. something we don’t want. »concludes Tudorache.