Abstract

In our modern era, Generative AI (GenAI) and Large Language models (LLM) are not buzzwords but rather the tools used to reshape our writing, learning and assessment experience in higher education. As such, even though there are evident advantages throughout our everyday use, from facilitating the learning support to language mediation and faster feedback, GenAI has several disadvantages as well. Specifically, it introduces significant issues in terms of reliability and fairness, most commonly including ambiguity of authorship, production of persuasive but often false and inaccurate claims, fabricated false citations and overall, deteriorating the process assessment validity as we know it. This means that the traditional assignment formats are starting to become outdated and unusable in both digital and physical education. At the same time, institutional responses do not state whether the pros outweigh the cons, but rather focus on detecting the GenAI or LLMs used to punish it or to take a more defensive action by stating false evidence in such research or providing AI and similarity statistics by detection tools and metrics. In this way, our article aims to provide an ethical and regulatory governance standpoint of GenAI in higher education. Analytically, firstly, we showcase the importance of university-developed strategies and policies for GenAI use for both teachers and students alike. Then, we address the elephant in the room, i.e., the necessary critical thinking needed to redesign, monitor and re-evaluate digital online assessments. Thirdly, we present governance principles and ethical concerns, and focus on an overview of the taxonomy of existing policy controls to help individuals and organizations monitor and organize their actions. Lastly, it is noted that our technical merit is providing 2 Tables regarding a detailed overview of policy controls across different governance domains and a comprehensive comparative policy control matrix covering Asia, the EU, Oceania and the USA. This is also supported by a threefold maturity-level model focusing on why policy controls matter that focuses on assessment integrity, fairness and data protection principles. Our framework can also extend to military education contexts (e.g., DoD, NATO, Air University) to ensure AI integration and mission-critical integrity, ethical decision-making and data security under operational constraints..

Keywords:Generative artificial intelligence; Higher education; Military education; Defense AI governance; Digital education; AI governance; Academic integrity and AI; Assessment design; Fairness AI assessment; Institutional policy; Data protection