EU's AI Act mandates labeling for AI-generated content, with hefty fines for non-compliance

The European Union's landmark Artificial Intelligence Act (AI Act) has entered a crucial enforcement stage, implementing transparency requirements that affect millions of AI applications both within and outside the bloc. Starting August 2, AI systems must clearly inform users about their nature, marking a significant step in regulating the rapidly evolving field of artificial intelligence.

Under the new rules, interactive AI systems, such as chatbots, must disclose that users are interacting with an AI rather than a human. Furthermore, any images, videos, or audio edited or generated by AI, commonly known as deepfakes, must be clearly labeled. These AI-generated or manipulated contents will also require machine-readable tags to facilitate easier detection, enhancing transparency and combating misinformation.

These obligations apply to both providers who develop and release AI systems and deployers who use these systems in their operations. This includes media companies publishing AI-generated news reports, banks using third-party chatbots, and recruitment firms employing AI for resume screening. The transparency requirements are immediately applicable to all AI systems within their scope, regardless of when they were launched.

Organizations failing to comply with these transparency mandates face significant financial penalties. Fines can reach up to 15 million euros or 3% of their global annual turnover, whichever is higher. Unlike many other provisions of the AI Act, these transparency obligations have no extended grace period and apply immediately. The EU aims to reduce deception, misinformation, and manipulation while safeguarding democratic processes and public trust.