General purpose AI systems (and particularly language models) are showing enormous potential for innovation but their development is also raising concerns over emerging risks. This article explores the regulatory concerns surrounding general purpose AI systems, especially focusing on the requirements outlined in the different amendments put forward by the Council and the Parliament for the proposed AI Act. Against the risk-based background of the regulation, the article analyses the two proposals and stresses the importance of addressing the risks associated with general purpose AI systems while promoting responsible use throughout the value chain.