EU's AI Act: SMEs Face New Rules, Fines for AI Use
The EU is set to introduce the AI Act, a comprehensive regulation for artificial intelligence in the economy. Medium-sized businesses (SMEs) using AI systems should take note, as these will be primarily affected, not just developers.
SMEs are the primary users of AI systems, often employing large language models and AI-integrated tools like Notion or Microsoft 365. Under the AI Act, these systems may fall under regulation, especially high-risk applications such as biometric identification, critical infrastructure, and medical devices. Companies must conduct risk assessments and document them, even for minimal risk.
The AI Act introduces four risk levels, with 'high' and 'limited risk' categories requiring specific obligations like documentation and transparency. From August 2025, transparency and documentation obligations become mandatory for generative AI models used in products or services. Companies must classify their AI systems according to risk and comply with transparency, risk management, and quality standards to avoid substantial fines (up to €35 million or 7% of annual turnover).
New documentation and transparency requirements include labeling AI-generated content and informing users about AI interactions. Companies not yet implementing the AI Act should start by taking an inventory of their used AI systems and determining their risk levels.
The EU AI Act is coming, and it will significantly impact SMEs using AI systems. Companies must prepare by understanding the risk levels of their AI applications, conducting risk assessments, and ensuring compliance with the new regulations to avoid substantial fines.
