The OECD today publishes guidelines for the responsible adoption of generative artificial intelligence in higher education: five policy areas to strengthen governance, quality and academic integrity in universities.

Generative AI is now a constant presence in the daily lives of students, faculty, and administrators of higher education institutions. This is confirmed by a recent OECD report, “Policies supporting responsible and systematic Generative AI adoption in higher education”, published as part of the OECD Education Spotlights series. The document photographs a phenomenon of very rapid growth and proposes, at the same time, a framework for a more structured governance, capable of accompanying the adoption of generative artificial intelligence without passively suffering its effects.

For an institution that, like MQE, closely follows the evolution of quality assurance systems in higher education, the OECD report offers valuable insights. This is not just a technological issue: the adoption of generative artificial intelligence directly affects issues central to institutional quality, such as academic integrity, data protection, equity of access and the reliability of content produced to support learning.

Widespread adoption: OECD data on generative AI

The numbers collected by the OECD give a clear picture. In the UK, the share of university students using generative AI tools rose from 66% in early 2024 to 95% in 2026. In Germany, surveys indicate an adoption rate of more than 90% among students. At the European level, 72% of students say they have used generative AI in the last three months, with a range ranging from almost 90% in Estonia to just over 50% in Turkey.

Academic staff are also turning to these tools, albeit at a more cautious pace: globally, 61% of faculty report weekly use, but only 17% consider themselves advanced users. The use remains concentrated on operational activities, such as writing texts, revising and preparing teaching materials.

In the face of this spread, however, institutional governance has not kept pace. A recent UNESCO survey found that only 19% of higher education institutions in the world have a formal policy on generative artificial intelligence, while a further 42% is still in the works. As a result, students and staff often find themselves dealing with sensitive issues such as data protection, copyright, and fair academic use policies individually.

The five policy areas identified by the OECD

Based on examples collected in member and partner countries, the OECD identifies five areas where systematic policy responses for the adoption of generative AI are emerging:

1. Guidelines for responsible use, with national frameworks developed by Ireland, Australia and the United Kingdom, among others.
2. Common approaches to compliance and purchasing, such as the Italian national agreement with OpenAI or the Australian collaboration with Anthropic.
3. Development of specific skills, through training programs carried out in Germany, Switzerland, South Korea and other countries.
4. Evidence collection and pilot projects, including the Australian GenAI Knowledge Hub and Dutch eduGenAI trials.
5. Development of generative artificial intelligence tools dedicated to the educational world, such as the Ethel teaching assistant of the ETH Zurich (ETH Zurich).

The report highlights how these initiatives share a common principle: the adoption of generative artificial intelligence gives positive results only when it is accompanied by secure infrastructure, clear policies, staff training and continuous evidence-based evaluation, rather than left to individual initiative alone.

The Maltese snapshot: adoption at European summits, policies still fragmented

The Maltese context also offers significant indications on the phenomenon described by the OECD report.

According to the recent special Eurobarometer 572 – The Digital Decade 2026, published by the European Commission, Malta is the first member state of the European Union for daily use of generative artificial intelligence:

• 38% of Maltese say they use tools such as ChatGPT, Google Gemini or Microsoft Copilot every day in their personal lives, almost double the European average, which stands at 20%.

The figure rises further if we consider professional and academic use:

• 39% of Maltese respondents say they use generative artificial intelligence on a daily basis at work or in studies, the highest percentage recorded in the entire Union.

The survey, conducted on more than 26,000 European citizens, including 518 residents of Malta, also shows that 86% of those who already use generative artificial intelligence have increased their use in the last year, with 55% speaking of a “significant” increase, also the highest value in Europe. At the same time, 84% of Maltese say they are in favour of regulating artificial intelligence to ensure its safety, even at the cost of introducing restrictions for developers.

On an institutional level, some Maltese universities and providers have already moved in this direction. The University of Malta has adopted a policy dedicated to the responsible use of generative AI, which encourages its use as a learning accelerator as long as it is declared in a transparent manner, while undeclared use in an evaluated work is equated to a case of cheating under the university’s evaluation regulations. MCAST has also developed, with the contribution of its Student Council, a policy document structured around five areas – alignment with the European AI Act, data protection, academic integrity, AI literacy and transparency – which allows the use of generative artificial intelligence in academic work as long as it is declared and not a substitute for the skills to be developed.

Some private providers, such as GBS Malta, have also introduced their own specific policies, with detection thresholds based on tools such as Turnitin. The picture that emerges is consistent with what has been observed internationally by the OECD: a very rapid and widespread adoption of generative AI, in the face of institutional responses developed individually by each entity, in the absence – at the moment – of a single and coordinated sectoral framework at national level.

Academic quality and integrity: the repercussions for universities

For higher education institutions, the adoption of generative AI is not just about everyday teaching, but directly affects internal quality assurance systems. The OECD report highlights risks related to the protection of personal data, the integrity of assessments, the unequal access between students and the reliability of the content generated, which in some cases can compromise the development of authentic critical skills.

These elements are intertwined with areas already familiar to those working in the field of higher education quality: programme design, student evaluation criteria, public information policies and document management. An institution that intends to adopt generative AI tools in its educational offer should therefore carefully evaluate how these tools fit into its quality management system, defining clear internal policies on permitted use, verification of content, training of teaching staff and protection of students.

The OECD also recommends ensuring equitable access to secure generative AI devices, connectivity, and tools, investing in staff training, enforcing strict privacy and security standards, supporting research into the educational impact of these tools, and maintaining meaningful human oversight at all times, along with alternatives that do not depend on generative AI.

MQE’s role: supporting institutes and universities in the governance of generative artificial intelligence

In this scenario, MQE can support further and higher education institutions operating or intending to operate in Malta in structuring an approach to generative artificial intelligence consistent with the quality assurance standards required by the Maltese regulatory framework. The support may concern, in particular, the drafting of institutional policies on the use of generative AI by students and teaching staff, to be integrated into the Internal Quality Assurance (IQA) documentation already required for licensing and accreditation purposes; aligning these policies with the assessment policy, the student handbook and existing disciplinary procedures, to avoid overlaps or inconsistencies; the definition of transparency criteria and declaration of the use of generative AI in the works evaluated, consistent with the practices already adopted by other Maltese universities; and an initial assessment of the risks associated with the use of these tools in the design of programmes, in the preparation of teaching materials and in the management of institutional information.

MQE can also support institutions in the preparation of training courses for teaching and administrative staff on the responsible use of generative AI, in the review of the monitoring and periodic review systems of the programs in the light of these new tools, and in the analysis of the degree of documentary preparation in view of any audits or external quality assurance cycles. It should be noted that the introduction of a policy on generative AI does not replace an overall regulatory assessment of institutional compliance and that each intervention must be calibrated to the specific situation of the entity: for this reason, MQE recommends a preliminary orientation session aimed at identifying priorities and any documentation gaps before proceeding with the drafting of internal policies or procedures.

Towards a structured governance of generative artificial intelligence

The main message of the OECD report is that the phase of the simple deployment of generative artificial intelligence is now over: the current challenge concerns the ability of higher education systems to integrate these tools in a way that enhances learning, without compromising academic integrity, privacy and equity. Effective governance requires explicit institutional choices, not simply progressive adaptation to a use already widespread among students and teachers.

For institutions operating or intending to operate in Malta, these issues are in addition to the obligations already laid down in the national quality assurance framework. Defining a structured approach to generative AI in time, consistent with its AQI policies, can represent an element of institutional solidity, as well as a credibility factor in the eyes of students, staff and supervisory authorities.