EU Introduces Stricter Transparency Laws in AI Act

The AI Act’s implementation will occur in phases over the next two years, allowing regulators to gradually enforce the new requirements while businesses adjust to their new obligations. However, the specific mechanisms for enforcing these rules are still under development. The newly established AI Office plans to release a standardized template for these disclosures by early 2025, following consultations with industry stakeholders.

This legislative move comes in the wake of a substantial increase in public engagement and investment in generative AI, fueled by the launch of OpenAI’s ChatGPT, a Microsoft-backed project. Generative AI applications, capable of producing text, images, and audio content swiftly, have captivated the public and investors alike, but they have also raised critical questions about data sourcing and copyright infringement. Concerns have been voiced over whether the use of copyrighted material, such as bestselling books and Hollywood movies, to train AI models without explicit permission from the creators constitutes a breach of intellectual property rights.

Read more: US Antitrust Chief to Prioritize AI Sector Amid Monopoly Concerns

One of the more controversial aspects of the AI Act is the requirement for organizations using general-purpose AI models, like ChatGPT, to provide “detailed summaries” of the training data. This transparency measure aims to address growing concerns about the ethical and legal implications of AI development. However, AI companies have expressed strong resistance to these demands, arguing that such disclosures would reveal trade secrets and potentially undermine their competitive edge.