Generative AI is already in education. The question is no longer whether schools will encounter it, but what kind of support they will need to use it well.
That is where GenAI4ED has been focusing its attention. Over the past year, the project has not only reviewed the wider landscape of AI in secondary education, but also worked directly with students, educators, and parents to understand what they actually need from a platform designed for this space. Across the project’s early research and co-design work, one message comes through very clearly: schools are not asking for more AI hype. They are asking for clarity, trust, and practical guidance.
That makes this moment especially important. Earlier GenAI4ED blog posts have already shown that education systems are under pressure to integrate AI, that academic integrity concerns are rising, and that schools need stronger ways to evaluate tools before adopting them. This new phase of work adds something more grounded: direct evidence of what stakeholders want to see in practice.
They want help, not just access
One of the most useful things GenAI4ED has found is that stakeholders are not simply asking for access to GenAI tools. They want help in making sense of them.
In the project’s co-design process, end-user requirements were gathered through multilingual questionnaires, participatory workshops, and expert validation. The purpose was not only to understand preferences, but to identify what would make a GenAI platform genuinely useful, safe, and practical in everyday school life. This early work was carried out collaboratively across the consortium, with particularly strong input in the stakeholder co-design and synthesis phase from The Cyprus Institute.
These insights did not come from a handful of impressions. They were shaped through a multilingual, role-adaptive questionnaire completed by 271 participants, and complemented by three focused workshops involving 24 participants in total. Together, these activities brought in views from educators, parents, and students, helping the project move from broad signals to more concrete design choices.
Across these groups, there was a remarkably consistent demand for structured support: tutorials, understandable guidance, transparent tool information, and simple explanations of how GenAI works. But the emphasis differed by role. Parents tended to focus on safety, oversight, and clarity. Educators emphasized classroom fit, reliability, and integrity. Students wanted credible recommendations and quick help finding what is actually useful for learning.
That is an important signal. It suggests that schools do not need a platform that simply lists tools. They need one that helps users judge them.
They want trust to be visible
Another strong insight is that trust has to be built into the user experience, not added later as a warning label.
The co-design findings pointed toward a hybrid platform model with a role-aware dashboard, a conversational assistant, a Training Hub, a Community Space, and a Tool Explorer supported by clear trust and privacy cues. The same work identified five core design principles: start with structure, keep guidance close, reduce friction, deliver training in context, and personalize carefully without overwhelming the user. Privacy-by-design, age-appropriate consent, accessibility, and scalability were treated as core requirements, not optional extras.
This makes the project’s direction especially relevant to current school debates. The most recent GenAI4ED platform blog also presents the platform as a safe, practical, and user-friendly environment, while the evaluation-focused article argues that schools need to look beyond technical performance and ask harder questions about pedagogy, privacy, fairness, and wellbeing.
In other words, GenAI4ED is moving in the same direction that schools increasingly need: away from “try it and see” and toward more supported, evidence-informed use.
They want evidence before enthusiasm
What makes this work especially valuable is that it does not stand alone. It builds on a wider body of project research that helps explain why these stakeholder needs matter so much.
The project’s earlier review work synthesized findings from 59 academic studies, an initial taxonomy and first-order evaluation of 28 GenAI tools, and 72 policy or guidance documents. Across that work, the same themes appeared repeatedly: people see real value in GenAI for personalization, engagement, content creation, and feedback, but they also worry about plagiarism, privacy, inaccuracies, dependency, bias, and the lack of practical guidance. AI literacy and training emerged as especially important across the board.
That is exactly why the co-design phase matters. It moves the discussion from a general diagnosis of risks and opportunities to something more actionable: what a useful support environment should actually look like for different users.
They want solutions that fit real school contexts
The same principle carries into the pilot studies.
Instead of designing one generic demonstration, GenAI4ED has developed three pilot pathways across Italy, Cyprus, and Greece, each reflecting different school realities. One focuses on varying levels of AI autonomy with younger students. Another explores subject-specific integration in STEM, Arts, and English. A third looks at broader cross-curricular use and wellbeing in daily platform use. Together, these pilots are intended to test not only whether the platform works, but how it works in real educational contexts, with different ages, subjects, school types, and support needs.
That matters for readers because it shows that GenAI4ED is not building around a single imagined classroom. It is building around the reality that schools differ, and that responsible GenAI adoption has to make room for those differences.
What this means for schools now
Perhaps the most encouraging message from this stage of the project is also the simplest one: schools already know what they need more of.
They need better guidance. They need clearer ways to evaluate tools. They need training that respects time and context. They need support for students, educators, and parents alike. And they need this to happen without losing sight of privacy, trust, fairness, and educational purpose
Author: Emmanouil Kritikos (The Cyprus Institute)