On Sunday, the European Union commenced the enforcement of its artificial intelligence legislation, signaling a shift towards stringent oversight and significant penalties for non-compliance.
Known as the EU AI Act, this pioneering regulatory framework officially took effect in August 2024. With the deadline for new restrictions on certain AI systems being met on Sunday, companies are now required to comply fully or risk penalties.
The rules include mandates for sufficient tech literacy among employees, and bans on AI applications considered to carry “unacceptable risk” to individuals, such as social scoring systems, real-time facial recognition technologies, and biometric tools that sort individuals based on sensitive attributes like race or sexual orientation. Additionally, “manipulative” AI technologies are also prohibited.
Companies found in violation of the AI Act may be subject to fines reaching up to 35 million euros ($35.8 million) or 7% of their worldwide annual revenues, depending on which figure is higher.
This marks a more severe penalty in comparison to those under the General Data Protection Regulation (GDPR), where infractions can result in fines up to 20 million euros or 4% of a company’s global turnover.
The size of penalties will be contingent upon the severity of the breach and the scale of the offending organization.
The current enforcement of the European Union’s AI Act represents just the first step in a succession of future regulatory measures, as the law is not yet in full effect. According to Tasos Stampelos, Mozilla’s head of EU public policy and government relations, the AI Act, though not without its flaws, is a crucial piece of legislation, largely focused on product safety standards.
However, the law has sparked discontent among several tech leaders and investors who argue that its stringent provisions could hinder innovation.
In mid-2024, Prince Constantijn of the Netherlands expressed his worries about Europe’s regulatory focus on artificial intelligence during an interview with CNBC. He highlighted that while regulatory clarity and market predictability are positive, achieving these in the rapidly evolving AI landscape poses significant challenges.
Meanwhile, others believe that Europe’s initiative to establish clear AI rules could serve as a strategic advantage.
Diyan Bogdanov, an engineering director at Bulgarian fintech company Payhawk, stated that while the U.S. and China are racing to develop the largest AI models, Europe leads in creating the most reliable ones.
He added that requirements in the EU AI Act concerning bias detection, periodic risk evaluations, and human oversight are not stifling innovation but rather setting standards for excellence in the industry.
