.png){kind=small}
On Wednesday, 13th March 2024, the European Parliament approved the Artificial Intelligence (AI) Act.
The regulation was agreed in December 2023 and aims to protect fundamental rights, democracy, the rule of law, and environmental sustainability from high-risk AI.
Here are some key elements:
AI Act: Bans
The Act bans AI applications that threaten citizens' rights, including:
Systems that categorize people based on sensitive characteristics (race, religion, or sexual orientation) using biometrics.
Creation of facial recognition databases by scraping images from the internet or CCTV footage.
Emotion recognition technology in workplaces and schools.
Social scoring systems.
Predictive policing based solely on profiling.
AI that manipulates people's behavior or exploits vulnerabilities.
Law enforcement exemptions of the Act
Biometric identification (like facial recognition) is generally banned for police. "Real time" biometric identification may be allowed in very specific situations:
They need a court order or other legal permission beforehand.
It can only be used for a short time and in a limited geographical area.
Examples: finding a missing person or stopping a terrorist attack.
Even after an event, police need a court order for investigations of serious crimes.
AI Act: Obligations for high-risk systems
High risk systems that may cause harm to health, safety, environment etc have specific obligations to eliminate risks such as:
Maintain logs
Be transparent and accurate
Ensure human oversight.
AI Act: Transparency requirements
General-purpose AI (GPAI) systems must meet certain transparency requirements, including compliance with EU copyright law.
The more powerful these models are, the more the additional requirements.
Artificial or manipulated images, audio or video content ("deepfakes") need to be clearly labelled as such.