EU Enacts Landmark AI Regulation: What You Need to Know -

Business

EU Enacts Landmark AI Regulation: What You Need to Know

The EU has adopted the seminal AI Act, a world-first, horizontal regulatory framework for artificial intelligence that introduces a risk-based approach to AI governance. Coming into force on August 1, 2024, the Act prohibits those AI applications whose risk is considered unacceptable and sets specific demands for high-risk systems. Provisions under the AI Act will be implemented in phases, with full application expected in August 2026, although the deadlines for some specific rules are sooner.

Key features of the AI Act

Risk-based categorization: This means categorizing AI applications into different categories of risk and attaching different compliance obligations to each category.
Unacceptable risk: AI systems having an unacceptable risk shall be banned, including social scoring by governments and manipulative AI.
High-risk: Applications include those that are used in job applications or critical infrastructure, where strict requirements under the law exist.
Limited/minimal risk: Systems, like chatbots, are normally subject to transparency obligations, meaning a user needs to be informed when he or she is interacting with an AI.