Universities need to take the AI debate beyond the campus

From internal AI readiness to the university’s societal responsibility

Across Europe, universities are moving from experimentation with artificial intelligence tools to strategies, guidelines and compliance frameworks. Early responses have focused on teaching, assessment, research integrity, data protection and administration. These are necessary starting points, but they also show how strongly the AI debate is still shaped by internal systems, institutional readiness and risk management.

Indeed, collaboration often appears late, indirectly or as an extension of other responsibilities. In Sweden, for example, the higher education sector has addressed AI literacy, access to tools, research use, teaching, examination and administration. However, collaboration across society has been much less visible in policy discussions, despite this being part of what universities are expected to do as part of their core missions.

A 2025 European University Association report on AI that serves universities’ needs and values rightly argues that adoption should be guided by university values, not technological possibility alone. If universities are to uphold critical thinking, academic freedom, responsibility and public trust, those values must also shape engagement beyond the campus.

For universities, AI strategy cannot stop at education and research. It must also ask how AI reshapes collaboration, trust and public value – and who should influence the choices that guide its use.

In an AI-shaped society, universities may need to act less as producers of final answers and more as part of society’s sensemaking infrastructure: helping actors interpret weak signals, test emerging knowledge and act with judgement under uncertainty.

This calls for more than individual AI literacy. It calls for collaborative AI literacy.

AI as infrastructure for collaboration in higher education

AI is already changing collaboration.

Universities’ external partners use AI to search, summarise and interpret research, while public agencies, companies and civil society organisations experiment with AI-supported analysis and services. Universities are part of a wider network of partners, funders, regulators, users and knowledge actors through which value is created, and in which research, talent, funding, regulation and public services are increasingly mediated by AI.

This creates opportunities. AI can synthesise stakeholder input, map knowledge gaps, translate research, identify partners and support scenario work. It may become part of universities’ collaboration infrastructure: helping institutions see, connect and learn across ecosystems. But this is also relational infrastructure, built on trust, routines, shared language and long-term relationships.

It also creates risks. AI may flatten uncertainty, reproduce bias, privilege actors with better data and tools, or make collaboration appear more neutral than it is. University knowledge may become more accessible, but also more vulnerable to misinterpretation when stripped from disciplinary, methodological or ethical context.

The real question is not whether universities can adopt AI efficiently. It is whether they can help shape an AI transition that strengthens knowledge, trust and democratic capacity. To do so, collaboration must move to the centre of university AI strategies – not as an afterthought, but as a condition for responsible adoption.

Towards collaborative AI literacy, strategy and implementation

Universities need to complement individual AI literacy with collaborative AI literacy: the shared capacity of universities and partners to understand, question and govern AI use in joint work.

The EU AI Act reinforces this by making AI literacy a regulatory concern for organisations that provide or deploy AI systems. But universities and partners also need shared judgement about when AI-generated synthesis is useful, when it is misleading, how AI use should be disclosed and how responsibility is distributed.

Collaborative AI literacy is not only technical. It is social and institutional. It requires spaces where researchers, teachers, professional services, students and partners examine real use cases together: an AI-supported policy brief, an evidence review, a partner-matching tool or a citizen dialogue. The aim is to learn where AI strengthens judgement, where it weakens it and where human responsibility must remain explicit.

Crucially, AI policies are not only instruments of control or compliance. They are part of the learning environment itself: shaping expectations, responsibility, values and experimentation. If universities want staff, students and partners to develop sound judgement, policies must be designed as learning infrastructures, not only as rulebooks.

This raises a governance question: who is in the room when university AI strategies are developed?

Universities must protect academic autonomy and should not outsource strategic judgement. Yet they cannot understand AI’s societal implications by looking only inward. Funders, public authorities, regions, municipalities, industry, cultural organisations, civil society and citizens all encounter universities through collaboration. Their experiences should inform responsible AI adoption, use and governance.

The same applies to the tools universities adopt. AI systems that support engagement should be designed with academia, industry, government and civil society, not merely for them. Otherwise, technical systems may define what counts as a relevant partner, signal or collaboration.

AI, digital sovereignty and the practical shift universities need to make

This is especially important when European universities collaborate with large AI companies, many based in the United States. Such partnerships offer tools, infrastructure and expertise, but raise questions about data, dependency, procurement, language coverage, academic freedom and European digital sovereignty.

If collaboration infrastructure rests on externally controlled systems, universities need to be clear about what they gain, what they give away, what dependencies they create and how their values are protected.

This does not mean turning AI strategy into stakeholder management. It means recognising that the university’s public role is relational. AI will influence how problems are framed, evidence selected, expertise communicated and decisions justified. These are questions about the relationships through which universities and society create, interpret and use knowledge.

Three practical shifts would help:

• First, university AI strategies should become ecosystem-aware, mapping risks and opportunities in collaboration, innovation, public engagement and policy advice as well as in teaching, research and administration.
• Second, AI policy and literacy initiatives should include the university’s collaboration infrastructure: collaboration offices, libraries, communication teams, innovation support, legal services and partnership platforms.
• Third, universities should create arenas for co-learning with external partners, where AI is tested in concrete collaborative tasks and evaluated against academic values and public purpose.

The European debate on AI in higher education rightly focuses on integrity, regulation and readiness. The next step is to connect these concerns to universities’ societal mission.