Artificial Intelligence and Knowledge Production: Copyright, Ethics and the Limits of Governance : Policy Analysis and Practitioner Perspectives

Sammanfattning

The integration of generative artificial intelligence into academic research and knowledge production has created unresolved challenges for copyright law, research ethics, and institutional governance. While scholarly and institutional attention to these challenges has grown rapidly, existing governance frameworks have remained fragmented, reactive, and narrowly focused on individual user behavior, leaving the structural dimensions of AI in knowledge production largely unaddressed. This study examines the above issues by employing a mixed-methods qualitative approach comprising policy analysis and semi-structured interviews. There are five core governance texts analyzed in this study: UNESCO Recommendation on the Ethics of Artificial Intelligence (2021), Elsevier's AI-Assisted Writing Policy, and Harvard University, the University of Oxford, and the University of Cambridge's guidelines. These texts were analyzed together with interviews carried out with seven experts involved in conducting academic research, writing news stories, publishing, translation, and creative writing – all domains that have in common their interest in issues of authorship, originality, and responsibility in the era of artificial intelligence. Three theoretical lenses were used to guide the analysis: the Social Construction of Technology, Responsible Research and Innovation, and the Political Economy of Communication. The results show that there is a strong normative agreement among all five texts in favor of human responsibility and transparency; however, no comprehensive model integrating both ethical and copyright implications of using AI technologies has emerged yet. Most importantly, none of these governance frameworks considers the issue of corporate ownership of AI infrastructures, epistemic inequality generated by AI learning algorithms, and power relationships between actors involved in the development of AI technology. Interview data confirms that practitioners across knowledge production fields are navigating AI governance without adequate institutional direction, developing personal ethical standards in the absence of clear policy guidance a condition this study describes as the privatisation of ethical responsibility. The study concludes that existing governance frameworks are normatively coherent but operationally insufficient. Effective governance of AI in knowledge production requires moving beyond individual responsibility frameworks toward structural approaches that address corporate power, technological dependency, and the intersecting legal and ethical dimensions of AI-assisted knowledge production.