Rudolph Angler A team of experts, appointed by the European Commission, is drafting a General-Purpose AI Code of Practice (CoP). The draft aims to guide companies in complying with the EU’s AI Act. Scheduled for publication in April, the code focuses on transparency, copyright, systemic risks, and mitigation measures.
The group of 15 European rightsholder organisations, including News Media Europe and the Federation of European Publishers, raised concerns about the draft. They claim it contradicts EU copyright laws. In a letter sent to European Commissioner Henna Virkkunen, they demanded changes to the draft.
Copyright Issues Highlighted by Rightsholders
The organisations warn that the draft Code of Practice undermines EU copyright rules. For example, it allows AI providers to make “reasonable efforts” to obtain lawful access to copyrighted content, instead of requiring proof of lawful access.
News Media Europe also criticized the code’s narrow interpretation of copyright rules for text and data mining. It argues that the draft creates “backdoor legislation,” preventing rights holders from verifying how AI models use their works.
The AI Act defines clear transparency and copyright-related obligations. However, the draft measures in the CoP are “insufficient” and “counterproductive,” according to the News Media Europe statement. Officials involved in the drafting process said the rules must be abstract to cover various industries, from music to text.
Next Steps and Global Implications
A third draft of the code will be published in mid-February. The Commission will decide whether to formally adopt it via an Implementing Act under the AI Act.
The AI Act will come into full effect this August, though bans on systems like facial recognition will begin on February 2.
Copyright issues involving AI extend beyond Europe. In the US, lawsuits over AI training data have already been filed. Comedian and author Sarah Silverman sued Meta and OpenAI in 2023, accusing them of using her memoir without permission to train their AI models.
The ongoing debates highlight the challenges of balancing innovation and intellectual property rights in the era of artificial intelligence.