New EU AI Code Requires Compliance with Copyright Laws

The European Commission has released a voluntary Code of Practice to guide AI developers in complying with the AI Act, emphasizing that companies must only use lawfully accessible, copyright-protected content for training. This framework requires adherence to existing legal obligations regarding transparency and security, ensuring that technological safeguards like paywalls are not circumvented. By aligning industry practices with EU copyright law, the code seeks to balance innovation with the protection of intellectual property rights. A central pillar of this initiative is the strict respect for web crawling restrictions, mandating that AI systems honor robots.txt files and other machine-readable protocols. This ensures that content creators’ decisions to limit access to their data are technically enforced during the scraping process. The code also establishes a dynamic list of known infringing sites and encourages dialogue between AI firms and rights holders to develop standards for expressing rights reservations, thereby fostering a more transparent and cooperative digital ecosystem. Relevance to open_data lies in defining the boundaries of data accessibility for public interest and commercial AI development. While promoting open access, the code clarifies that "open" does not mean "free from copyright constraints," necessitating rigorous due diligence in data sourcing. It introduces mechanisms for grievance redressal, ensuring that rights holders have a voice in how their data is used. This regulatory landscape forces data providers and AI developers to establish clear, lawful channels for data extraction, impacting how open data initiatives can be scaled without infringing on legal rights.

Source: medianama.com
Published on 2025-07-12

New EU AI Code Requires Compliance with Copyright Laws

Related news