On March 13, the European Parliament of the European Union voted into law the world’s first binding law on artificial intelligence (AI) titled “Laying Down Harmonised Rules on Artificial Intelligence (Artificial Intelligence Act)” to put forth a flexible coordinated approach regarding the “human and ethical” implications of AI. The stated objectives of this new regulation comprise ensuring that AI systems are safe and respecting the existing laws that protect the fundamental rights of citizens. Further, the act will aim to make sure AI investments are legally sound, enforce AI laws effectively, and develop a single market for AI applications. The European Parliament stated its legal basis for this law is through Article 114 of the Treaty on the Functioning of the European Union.

Following 89 preambles, the law’s Title I sets forth general subject matters: the law will set “harmonised rules” for the market as it will implement more AI systems; it prohibited certain AI systems; created specific requirements for “high-risk” AI systems; created transparency rules regarding emotion-recognition or biometric systems; and created rules on marketing surveillance. Title II prohibited AI practices and Title III covered “high-risk” AI systems. Prohibited AI systems include distorting someone’s behavior or vulnerabilities, evaluating trustworthiness with possibly a social score, or using “real-time” remote biometric identification systems for law enforcement (unless searching for either victims of a crime or missing children, among others). Title III on “high-risk” AI systems defined “high-risk” systems as those that pose fundamental risks to the rights of individuals, specifically to the health and safety of a citizen, among others.