Artificial Intelligence in Research: The Role of Central Regulation and University Autonomy.  Digitalisation – particularly through Artificial Intelligence (AI) – is profoundly reshaping society and public institutions, giving rise to new forms of power on a global scale. Since AI is not a natural resource, the ability to develop high-performing artificial intelligence systems depends on research and innovation activities and thus constitutes a form of such power, which is why both the public and private research sectors are competing for leadership in this field. In this scenario, universities, as natural hubs of innovation, stand at the forefront of the digital transition. First, they are actively involved in the research on AI systems, second, from a different yet complementary perspective, they are increasingly integrating AI into research activities, to accelerate and enhance scientific output. However, despite the significant attention devoted by legal scholars to the topic of AI, its twofold impact on academic research remains largely unexplored to this day. The present analysis will therefore examine how (i) the European legislation – namely the AI Act, GDPR, Data Act, and Data Governance Act – is regulating research and, especially, academic research for the development of AI, and (ii) how the European legal framework is disciplining the use of AI in academic research. To do so, two main dimensions will be investigated. First, the impact of EU regulation regarding the research and development of AI systems by academia will be assessed, with a focus on identifying exceptions to the binding rules laid down within the legislative framework. Second, it will explore how the legislation is regulating the use of AI in conducting research, especially by universities, seeking also to understand if academic institutions are beginning to regulate and structure the responsible use of AI in their own research environments. To better contextualise the European approach, a comparative perspective with the United States will also be introduced. Given the relative absence of a comprehensive regulatory framework for AI in the U.S., this comparison will shed light on the different institutional strategies adopted across the Atlantic to assess whether a “Brussels Effect” is beginning to emerge in the governance of AI in academia.