Policy Analysis: The EU AI Act Is Here. Are Universities Ready?

Policy: European Union Artificial Intelligence Act

Policy Milestone: Major implementation phase (Article 50) begins – 2nd August 2026

Policy Overview:

Universities are increasingly using AI across teaching, assessment, research and administration, making it increasingly important to consider what the EU AI Act could mean for higher education.

Today (2nd August 2026) marks an important milestone in the implementation of the European Union's Artificial Intelligence Act. Naturally, much of the media reporting of the legislation has focused on technology companies and AI developers. However, its implications extend far beyond the tech sector and should be an important consideration for universities. And, whilst some recent headlines have focused on the postponement of some “high-risk” AI obligations post-2026, this should not be interpreted as a delay to the Act as a whole.

Many important provisions of the act are coming into play today, and Universities should consider these carefully. Specifically, Article 50 now introduces transparency obligations for certain uses of AI at both an institutional and individual level. These rules for transparency are designed so that people know when they are interacting with AI or consuming AI generated content. 

So, that Chatbot may be an obvious inclusion, but what about a lecturer that has used some AI to prepare next weeks lecture slides? This is a €15 million question, as this is the possible penalty for breaching Article 50! Though, to calm you down a bit, we should point out this is one of the ‘ceiling’ figures and not an automatic penalty notice!

However, there is yet more to deal with....

The remaining obligations relating to many high-risk AI systems will be introduced in phases between December 2027 and August 2028.

Supporting Media Review:

Despite the significance of the EU AI Act for the world of academia, relatively little has been written about its implications with a specific regard towards universities. Most media coverage has concentrated on the impact of the legislation on AI developers, large technology companies and regulatory enforcement.

One notable exception is a recent Times Higher Education article by Seher Asaf, which examines how the legislation could specifically affect universities. This article was published before the EU confirmed the delay of some of the Articles in its AI Act, but given that the delay is merely a reprieve, the contents of the article remain highly relevant.

Drawing on comments from Thomas Jørgensen, Director of Policy Coordination and Foresight at the European University Association, the article suggests that universities may need to reassess some of the AI practices they have adopted since the rapid emergence of generative AI. Particular attention is given to the use of AI in student assessment. Jørgensen notes that AI-assisted assessment may raise compliance questions under the AI Act and observes that further guidance from the European Commission's AI Office will be important in clarifying how the legislation should be interpreted in practice. Rather than concluding that current approaches are unlawful, the article highlights an area where uncertainty remains and where universities should pay close attention to future developments.

The article also argues that universities may need to move beyond developing AI guidance for staff and students towards establishing more comprehensive governance frameworks. It further raises broader strategic questions about Europe's reliance on large language models trained predominantly on US datasets and the potential role of European-developed AI models in supporting linguistic and intellectual diversity within higher education.

In our view, the article provides one of the first substantive discussions of the EU AI Act from a higher education perspective and offers a valuable starting point for considering its potential implications for universities.

We will be keeping a close eye on this topic, so you can expect more information in future blog posts as the implementation date of the “high risk” articles of the EU AI Act approaches.

Why this is Important for A4-AI (Academics for AI):

The implementation of the EU AI Act represents an important moment for universities because it signals the beginning of a new more regulated phase in the relationship between universities and their usage of artificial intelligence.

Over the past two years, universities have understandably focused on practical questions surrounding generative AI: How should students use AI? How should assessment evolve? How should academic integrity be maintained? The EU AI Act introduces another dimension that deserves equal attention: AI governance and regulatory compliance.

While the detailed impact on higher education will continue to evolve as further guidance is published, universities may increasingly need to consider not only whether AI should be used, but also how it should be governed, documented and managed within institutional processes. However, governance is unlikely to be solely an institutional issue and universities may also need greater visibility of how AI is being used by individual academics in teaching, assessment and course development. After all, how much visibility do universities currently have over how individual lecturers are using tools such as ChatGPT to prepare teaching materials or assist with the grading of student assessments?

And, the EU AI Act is not limited in its relevance to EU universities.  As Article 2 of the Act states; “This regulation applies to (a) providers placing on the market or putting into service AI systems or placing on the market general-purpose AI models in the Union, irrespective of whether those providers are established or located within the Union or in a third country (c) (c) providers and deployers of AI systems that have their place of establishment or are located in a third country, where the output produced by the AI system is used in the Union”

As a final causal point of observation, it is notable, that as universities get to grips with the challenges of AI regulatory compliance, there is the scope to learn first hand from the experience and transfer this to the teaching of these important skills that some students themselves will have to take into the workplace once graduating.

Questions for Academia

Does your university have an AI governance framework that is capable of evolving alongside emerging regulation?

Should institutions begin reviewing how AI is used in teaching, assessment and student support in light of the EU AI Act?

Is there an opportunity for universities to learn from their own experience of AI regulatory compliance and transfer this to the teaching of such subjects to students?

The EU AI Act is still in the early stages of implementation, and many questions for higher education remain unanswered. As further guidance is published and universities adapt their policies and practices, A4-AI will continue to analyse the implications for teaching, assessment, research and institutional governance.

• Read the EU AI Act

• Read ´The Looming EU AI act could force universities to ‘change everything’ by Seher Asaf in the Times Higher Education