Starting August 2, European Union citizens will be required to be informed when they are interacting with AI systems or viewing content generated or altered by the technology. The new transparency regulations are part of the bloc’s Artificial Intelligence Act, which aims to reduce deception and manipulation while helping people make informed choices about AI use. The law mandates that AI-generated content, such as deepfakes, chatbots, and AI-driven complaint hotlines, be clearly labeled to ensure transparency. Companies using AI for scheduling, correspondence, or contract negotiations must also declare their use of the technology. Failure to comply could result in fines of up to €15 million or 3 percent of global annual turnover, whichever is higher.
The AI Act applies to a wide range of entities, including model developers and companies like Spotify and Adobe, which use AI in their services. The European Commission’s newly established AI Office will monitor compliance, ensuring that even less obvious uses of AI, such as AI-powered recommendations or editing tools, are properly labeled. The law mirrors the EU’s privacy-focused General Data Protection Regulation (GDPR), which reshaped how companies handle personal data and introduced widespread consent pop-ups. However, critics warn that the new law could lead to information overload, with excessive labeling potentially causing 'banner blindness' and diminishing the meaningfulness of AI content labels.
The EU has allowed a transitional period until December for AI model providers to label synthetic media in a machine-readable format, acknowledging the technical challenges of implementation. Despite delays in producing key compliance guidelines, the law’s sweeping reach is expected to significantly impact companies. Legal experts caution that enforcement may not be uniform across the bloc’s 27 member states, as they are at varying stages of setting up their supervisory frameworks. Nonetheless, the regulation is likely to be a 'gamechanger' for businesses, forcing them to either label AI use transparently or revert to non-AI systems.
Source: wired
