The classification of educational AI systems as “high risk” by the European AI Regulation (Regulation EU 2024/1689) redefines design constraints for any adaptive learning platform, automated assessment, or algorithmic guidance. This classification requires designers to rethink the very architecture of their tools, far beyond simply adding compliance features. Technological design in education is no longer just about user experience; it is a regulatory obligation.
Regulatory Constraints of the AI Act on the Design of Educational Tools
AI systems used for student admission, guidance, assessment, and monitoring fall under the “high risk” regime of the AI Act. We observe that most edtech solution providers have not yet integrated this reality into their product design cycle, even though the majority of obligations will apply in December 2027 for the education sector, following a postponement initially scheduled for August 2026 via the “Digital Omnibus” package (Regulation EU 2026/1744).
Practices already prohibited since February 2025 include emotion recognition in school contexts, social scoring applied to students, and manipulative techniques. The integration of compliance with this framework in the programs published by mitxdesigntech.org reflects an early consideration of these requirements in educational design research.
Specifically, any tool that grades, ranks, or guides a learner through an algorithm must provide detailed technical documentation, a risk management system, and mechanisms for human oversight. For a product designer, this means that interfaces must make algorithmic decisions explicit, not hide them under a layer of gamification.
AI Literacy: A Prerequisite that Changes Educational Design
Since February 2, 2025, the AI Act requires organizations deploying AI systems to train their staff in critical mastery of artificial intelligence. In the education sector, this obligation affects teachers, administrative staff, and, by extension, the learners themselves.
We recommend integrating this constraint directly into the design of platforms. A LMS that deploys adaptive learning without providing an explanatory module on how its recommendations work does not meet the literacy obligation. Design must embed algorithmic transparency as a native component, not as supplementary documentation.
The implications for designers are structural:
- Every algorithmic recommendation displayed to the learner must be accompanied by an accessible explanation of the criteria used, without excessive technical jargon
- Teachers must have a readable dashboard that outlines the scoring or grouping logic, with the possibility of manual correction
- Internal training pathways within institutions must include modules on the functioning of the AI tools they use daily
A well-designed tool makes AI literacy natural. A poorly designed tool makes it impossible, regardless of the associated training program.
Experience Design and Adaptive Assessment Systems
Adaptive assessment concentrates the tensions between pedagogical innovation and regulatory compliance. Platforms that adjust the difficulty of exercises in real-time use predictive models that, according to the AI Act, fall under high risk as soon as they influence the learner’s educational path.
The UX design of these systems must make the decision-making process visible without burdening the learning experience. We observe two dominant approaches in recent platforms. The first is to display an algorithmic confidence indicator next to each assessment, allowing the teacher to weigh the recommendation. The second integrates mandatory human validation points before any guidance decision.
The second approach is preferable in the current regulatory context. It ensures the human oversight required by the AI Act while maintaining the fluidity of the learner’s journey. The trap to avoid: designing human intervention as a mere checkbox. Supervision must be substantive, not cosmetic.
Edtech Funding in Europe and Development Priorities
The European edtech market is undergoing a phase of restructuring. Investments are focusing on solutions capable of demonstrating their compliance with the emerging regulatory framework, which is directing development priorities towards algorithmic documentation and system auditability.
This trend is changing the hierarchy of skills sought in product teams. Profiles combining expertise in interaction design, understanding of AI architectures, and knowledge of European digital law are becoming the most sought after. Edtech startups that hire only ML engineers without specialized compliance designers are taking a measurable regulatory risk.
Innovation in educational technology is now played out on data governance, not just on the sophistication of algorithms. Institutions selecting their digital tools are increasingly evaluating the quality of the compliance documentation provided by the publisher, alongside the educational functionalities.
What This Changes for Design Teams
Edtech specifications now incorporate traceability requirements from the specification phase. A prototype that does not provide an audit log of algorithmic decisions will be dismissed even before user testing. Design thinking applied to education must integrate the regulator as a full stakeholder in the design process.
The European regulatory framework, despite the postponement to December 2027, is already shaping technical choices and design trade-offs for educational platforms. Teams that wait for the deadline to adapt their tools are accumulating a compliance debt that will be costly to resolve. Designing for transparency from the start is cheaper than adding it later.
