EU förbjuder AI-system och inför böter på upp till $35 miljoner euro.

februari 3, 2025

Not marking your AI image as AI generated can get you in trouble

As of February 2, 2025, the European Union (EU) has begun enforcing its Artificial Intelligence Act (AI Act), marking a significant step in regulating artificial intelligence technologies. This legislation prohibits AI systems deemed to pose an “unacceptable risk” to individuals’ safety, livelihoods, and rights.

Understanding the AI Act’s Risk Categories

The AI Act employs a risk-based framework to classify AI systems into four categories:

Minimal Risk: These systems, such as email spam filters, are largely unregulated due to their low potential for harm.
Limited Risk: Includes applications like customer service chatbots, which are subject to light transparency obligations to inform users they are interacting with AI.
High Risk: AI systems used in critical sectors—such as healthcare diagnostics or autonomous vehicles—fall into this category. They must adhere to stringent compliance measures, including mandatory risk assessments and human oversight.
Unacceptable Risk: AI applications that are outright banned due to their potential to cause significant harm.

Prohibited AI Practices

The AI Act specifically bans several AI practices considered to have unacceptable risks, including:

Social Scoring: Systems that evaluate individuals based on personal characteristics or behaviors, similar to China’s social credit system.
Manipulative Techniques: AI designed to subliminally or deceptively manipulate individuals’ decisions, undermining their autonomy.
Predictive Policing Based on Appearance: AI that attempts to predict the likelihood of individuals committing crimes based solely on their physical appearance.
Emotion Recognition in Sensitive Environments: AI systems that infer emotions in workplaces or educational institutions without a valid medical or safety justification.
Biometric Surveillance: The use of real-time biometric identification systems, such as facial recognition, in public spaces for law enforcement purposes, with certain exceptions for specific threats.

Implications for Businesses

Organizations operating within the EU, or those providing AI systems used in the EU, must ensure compliance with the AI Act to avoid substantial penalties. Non-compliance can result in fines up to €35 million or 7% of the company’s annual global turnover, whichever is higher.

While the ban on unacceptable risk AI systems is now in effect, other provisions of the AI Act will be phased in over the coming months and years. Companies are advised to stay informed and proactively adapt their AI strategies to align with the evolving regulatory landscape.

The enforcement of the AI Act underscores the EU’s commitment to fostering a safe and ethical AI environment, setting a precedent that may influence global AI regulations.