Europe may officially pass the AI Act into law. The legislation focuses more strictly on AI systems that pose a high risk to the user. The manufacturer must start meeting transparency requirements for these systems, among other things.
The European Council approved the AI Act. The Council was the last party through which the bill still needed to pass. Therefore, it is now possible to officially convert the AI Act into legislation.
Risk approach
Under the AI Act, all AI systems available in Europe are assigned to a risk category. Higher risk has stricter rules attached to it, which should protect EU residents from high-risk systems. Systems used for defence and research purposes are not subject to the legislation.
High-risk systems are subject to transparency requirements and additional security procedures. The transparency policy requires developers to prepare technical documentation, comply with European copyright law and write a summary of training data.
Fines
Violations are punishable by monetary fines. The amount to be paid is determined based on the company’s annual global turnover or is a fixed amount, whichever is higher. Start-ups are fined more leniently.
Tip: AI Act: Europe is blind to the law’s innovation problems