What simply occurred? European Union lawmakers on Wednesday voted overwhelmingly to enact a landmark law governing artificial intelligence. Called the AI Act, the brand new regulation was permitted by 523 votes to 46 and is predicted to change the way in which AI is utilized by authorities companies and companies within the EU.
In a press launch following the vote, the European Parliament stated that the brand new law is the first-ever authorized framework on AI, and will do a balancing job of addressing the dangers associated to the use of the rising know-how whereas positioning Europe to play a number one function in its improvement globally. The lawmakers additionally acknowledged that the Act was designed to safeguard most of the people from dangerous uses of artificial intelligence whereas offering AI builders and organizations with binding tips on how to use it in a non-intrusive method.
According to the EU, the AI Act will safeguard the elemental rights of European residents, in addition to protect democratic values and environmental sustainability from “high-risk AI.” As half of the plan, the brand new law bans sure AI functions that the lawmakers consider threaten residents’ rights. These embody “biometric categorization methods based mostly on delicate traits and untargeted scraping of facial photographs from the web or CCTV footage to create facial recognition databases.”
Other functions within the “banned” record embody “emotion recognition” in workplaces and colleges, and social scoring. To prevent a dystopian future showcased in motion pictures like Minority Report, the law additionally categorically prevents predictive policing when it’s based mostly solely on profiling an individual or assessing their traits. Finally, AI that “manipulates human habits or exploits folks’s vulnerabilities” can be barred below the brand new law.
However, as with most civil rights and privateness legal guidelines, there are some exemptions for law enforcement. While the use of AI for policing has largely been forbidden within the new law, officers are permitted to use real-time biometric identification methods in some conditions if “strict safeguards” are met. For instance, police within the EU can use AI in focused searches for lacking folks or to prevent a terrorist assault, offered they get prior judicial or administrative authorization.
Coming to general-purpose AI software program, the law clearly stipulates that it should meet sure transparency necessities, together with “compliance with EU copyright law and publishing detailed summaries of the content material used for coaching.” The Act additionally specifies that AI-generated photographs or manipulated audio and video content material, in any other case generally known as “deepfakes,” have to be correctly labeled as such.
