Landmark AI Regulation Agreement Reached by European Union

By: Isha Das

A provisional agreement on comprehensive rules for the use of artificial intelligence (AI) systems has been reached by the European Union (EU) negotiators. This makes the EU the global pioneer in AI regulation space. The rules will not only advocate for safe and reliable AI innovation but will also safeguard the citizens' basic rights.

The AI regulation follows a risk-based approach distinguishing between high-risk and low-risk AI systems. High-risk systems will be obligated to follow stringent rules, whereas low-risk ones will be subjected to transparency obligations. For instance, AI-generated content is required to be disclosed, which will empower users to make a knowledgeable decision.

Apart from a risk-based approach, the other measures introduced include mandatory human oversight and data governance for high-risk AI systems; promoting transparency and explainability of AI inner workings; and proactive risk identification and mitigation. EU's agreement will outright ban certain AI functionalities, such as facial recognition in public spaces, AI that manipulates human behaviour, emotional recognition systems, and social scoring systems. Some exceptions will exist for law enforcement agencies.

Violating the AI applications laws can attract a penalty up to 7% of the company's global revenue. The legislation encourages the testing of innovative AI systems under real-world conditions with suitable modulations. Other major countries like the U.S., U.K., and Japan can look towards these regulations as they also attempt to form their legislation for AI system's management.