The EU hammered out a deal on the comprehensive regulation of AI. Thierry Breton, commissioner for internal market of the European Union, took to the X social media platform to hail the AI Act as harnessing the potential benefits of the rapidly advancing technology, while mitigating possible risks.

The draft agreed by EU lawmakers would ban:

▪️Use of real-time surveillance and biometric technologies, with three exemptions;

▪️Untargeted extraction of facial images from the Internet or CCTV footage for setting up facial recognition databases;

▪️Emotion recognition in workplace and educational institutions;

▪️Social scoring – determining an individual’s social behavior – based on social performance or personal characteristics;

▪️AI systems able to manipulate human behavior to bypass their free will;

▪️AI used to exploit the people’s vulnerabilities based on age, disability, etc.

Special obligations, such as “fundamental rights impact assessments,” are set out for AI “high-risk systems,” which could potentially cause harm to health, safety, fundamental rights, environment, and rule of law.

General-purpose AI systems will adhere to “transparency requirements,” with obligations including disclosure of data used to teach the machines.