New Tools Emerge for Automated EU AI Act Compliance
Researchers have introduced "Compliance Cards," an automated system for analyzing and documenting adherence to the EU AI Act across complex supply chains. This approach allows SaaS vendors to provide clients with auditable, real-time compliance status. The tooling emerges as the EU AI Act moves to mandate clear labeling of AI-generated content, with non-compliance fines reaching up to 7% of global revenue.
- The EU AI Act is being implemented in phases; rules banning certain "unacceptable risk" AI practices began applying as of February 2025, with the majority of rules, including those for high-risk systems and transparency, becoming enforceable by August 2, 2026. - The regulation categorizes AI systems into four tiers of risk: unacceptable (which are banned), high-risk, limited-risk, and minimal-risk. Prohibited applications include social scoring systems, cognitive behavioral manipulation, and most real-time remote biometric identification in public spaces. - High-risk AI systems, such as those used in critical infrastructure, employment, or law enforcement, face stringent obligations. These include requirements for continuous risk management, high-quality data governance to prevent bias, detailed technical documentation, human oversight, and a high level of cybersecurity and robustness. - The penalty structure for non-compliance is severe, with fines for using prohibited AI systems reaching up to €35 million or 7% of a company's total worldwide annual turnover, whichever is higher. - Providing incorrect or misleading information to authorities can result in fines of up to €7.5 million or 1% of global turnover. - By August 2, 2026, deployers of AI systems that generate or manipulate image, audio, or video content constituting a "deepfake" must disclose that the content is artificial. Similar transparency rules apply to AI-generated text published to inform the public on matters of interest, unless it has undergone human editorial review. - To facilitate compliance with transparency rules, the European Commission is developing a Code of Practice, expected by mid-2026, which will provide guidance on labeling, watermarking, and other technical measures to identify AI-generated content.
