Agreement on Landmark EU AI Act Reached
The European Union has agreed on principle on the world's first comprehensive regulation on Artificial Intelligence, the AI Act. Initially introduced a few years ago, European legislators have been prioritizing the bill over the past several months as the world reacts to the proliferation of AI tools like ChatGPT.
The AI Act takes a tiered approach to AI development, regulating technologies rather than specific use cases.
- Unacceptable risk, such as social scoring and mass surveillance technologies
- High risk, such as AI developed for law enforcement purposes
- Limited risk, such as chatbots and deepfake technologies
- Minimal risk, such as spam detection
As you go up the risk scale, requirements will become harsher; minimal risk systems do not carry additional obligations, limited risk has transparency obligations, high risk systems must complete conformity assessments, and unacceptable risk systems are prohibited.
The text of the law is not yet fully available, but penalties could incur fines of up to 7% of annual worldwide turnover. For more on our initial reactions to the historic agreement, click here.
