As AI is expanding its role in our everyday lives, governments around the world are in a rush to enact their own set of regulations. The goal is to mitigate AI risks for citizens and pave the way for trustworthy AI development. The US first introduced the AI Bill of Rights as a set of principles for safe AI deployment. It was then followed by the EU with the AI Act.
While AI regulations in the United States are still changing hands between lawmakers, the EU’s Parliament and Council negotiators reached a vital agreement to regulate AI on the continent.
According to The Washington Post (via Engadget), a 72-hour debate between negotiators resulted in a historic agreement. The agreement is said to be the most comprehensive of its kind. The deal was reached while nations like France, Germany, and Italy were worried that enacting new laws might prevent EU members from developing cutting-edge AI technologies.
The EU takes a big leap forward in regulating AI in the continent
The new legislation basically proposes a set of guidelines for organizations that want to rely on AI for their work. For example, proposed regulations prohibit entities from using AI to design biometric categorization systems. Such systems could be used to target civilians based on their philosophical beliefs, political, religious, etc.
Additionally, using AI to create facial recognition databases and exploit people’s vulnerabilities is banned. According to the EU’s new legislation, designing AI systems to manipulate human behavior, designing social scoring systems, and launching emotion recognition in the workplace are other examples of banned AI applications.
The legislators have considered some exemptions for law enforcement agencies to use biometric identification systems (RBI). The RBI would be used in cases such as looking for abduction, trafficking, and sexual exploitation victims. Other uses are finding criminals and preventing a terrorist attack. Most details about the new legislation are still undercover. However, the MEPs also mandated high-risk and general AI systems to a set of rules for protecting civilians’ rights.
2023-12-14 15:07:31