With the increasing need of AI governance, the European University Association sheds light on essential key points to consider for universities.
As people gain larger access to generative AI, the use of AI is becoming a widespread practice among many individuals. Increasingly, institutions are adopting strategies and frameworks, such as the EU AI ACT and the AI in Science Strategy (see SwissCore article), to implement the best practices. In this context, universities are also concerned about the use of AI within its institutions and among its students and researchers. As they can influence society’s stance towards the use of AI, their approach should be thoughtfully designed. The European University Association (EUA) report on “Adopting AI that serves the needs and values of universities” highlights that the difficulty of AI governance lies in the fast-paced advancement of technologies. Indeed, the rapid change of the sector makes it difficult for universities to predict and anticipate their needs, set clear priorities and allocate resources. Therefore, the report aims to serve as a starting point for reflection, as well as providing a guideline for universities faced with the pressure of adopting AI that best serves their needs. The report sheds light on 5 key areas: i. ethics, principles and values, ii. institutional strategies, iii. training, iv. regulation and compliance, v. sustainability and societal impacts.
While AI contributes in efficiency, it is important for universities to first identify the added value of the tools and to be aware of the risks and limitations of the system. One way to ensure that AI adds value is to use it responsibly and to consider the ethical aspects. Transparency, privacy, just representations and preservation of academic values and integrity are essential ethical points to take into account. The EUA report emphasises on following a human-centric approach – meaning that the AI is a tool that enhances human activities but do no control them – as well as focusing on training the users on the system, its limitations and risks.
When it comes to institutional strategies of universities, it is essential to adopt sustainable approaches that ensure long‑term impact. The core values should be based on the institutional missions. Here again, AI should be a tool that enhances university values and respectfully serves its missions, while enabling room for experimentation.
The report also highlights the importance of training as AI will be embedded in many different areas which encompass from science to social topics. This makes AI literacy an essential skill both for future academic and professional opportunities. For instance, the importance of promoting digital awareness is also apparent in the EU Digital Competence Framework for Citizens. In the field of research, the use of AI in research process is already becoming an integral part of it. Thus, universities should ponder on adopting AI guidelines that best meet their goals and values, while enhancing the benefits of AI in supporting research practices. Trainings will allow to identify the opportunities and the limitations, and universities could act as important avenues for exploration and discussion on the topic.
The biggest challenge in the adoption of AI relies largely on the implementation of the EU AI Act and the compliance to it. The AI Act is grounded in trust and regulates AI according to different levels of risk. However, determining the level of risk of AI practices is not always clear and lays in a grey zone. As the AI Act envisions a system of national bodies overseeing the implementation and compliance, this could be an opportunity for universities to closely collaborate with these institutions to share knowledge.
Finally, environmental and social impacts are inherent to the use of AI. These impacts should be carefully thought and universities have a role to play: raising greater awareness on the consequences of using AI, promoting social justice and initiatives mitigating social disadvantages, as well as supporting exchanges.
For Swiss universities, as they are also in the transition towards a greater adoption of AI, it seems necessary to embrace a flexible and exploratory stance, emphasise on training and foster collaboration between different stakeholders.