Generative AI as a Tool for Inclusive and Accessible Higher Education

Digital transformation is reshaping higher education. Among the most powerful innovations is Generative Artificial Intelligence (GenAI), which includes tools such as large language models, advanced translation engines, and text-to-speech systems. When designed and used responsibly, GenAI can help make tertiary online education more inclusive and accessible for all learners.

Why inclusion matters

European policy strongly emphasises that digital learning must be open to everyone. The Digital Education Action Plan (2021–2027) [1] highlights inclusion and accessibility as core goals, while UNESCO [2] and the OECD [3] underline the importance of using GenAI to enhance—not replace—human-centred teaching. These frameworks set the stage for universities and projects like Erasmus+ initiatives to experiment with AI responsibly.

How GenAI supports accessibility

Multilingual support: GenAI-powered translation and multilingual interfaces allow international students to learn in their preferred language. This lowers barriers for students in cross-border online programmes [4].
Simplifying and adapting learning content: AI can rephrase complex academic texts, generate plain-language summaries, or create alternative representations of concepts. This is especially useful for students with dyslexia, ADHD, or limited language proficiency [5].
Captioning and transcription: Automatic speech recognition enables live captions and transcripts for video lectures. These tools support deaf and hard-of-hearing students and help non-native speakers follow classes more easily.
Personalised study support: AI assistants can generate study plans, summarise readings, or break tasks into smaller steps. This scaffolding helps students who need more structure, such as neurodivergent learners.
Tutoring and feedback: Chatbots can provide immediate answers, explanations, and feedback. While not a substitute for educators, they offer additional support for students who may otherwise struggle to access help [6].

Risks and challenges

Of course, GenAI is not a silver bullet. Several risks must be addressed:

Bias and misinformation: AI systems can reflect and amplify bias or generate incorrect answers. Human review remains essential [2,3].
Digital divide: Not all students have equal access to devices, connectivity, or AI skills. Institutions must provide equitable alternatives [7].
Accessibility gaps: Automated captions and translations are not always reliable enough for critical teaching. Human correction is often needed.
Privacy and data protection: Under the GDPR, universities must carefully manage how student data is used in AI systems.
Academic integrity: AI-generated text raises concerns about plagiarism. Clear guidelines and authentic assessment design are better solutions than unreliable AI-detection tools [8, 9].

European standards and regulations

Institutions deploying GenAI in education need to align with various European rules and standards that are already implemented in online learning. Some of these rules and standards are:

WCAG 2.2 [10] and EN 301 549 [11]: Accessibility standards ensuring online learning materials are usable by all.
European Accessibility Act (from June 2025) [12]: Legally requires accessible digital products and services, including educational platforms.
EU AI Act (2024) [13]: Education is considered a “high-risk” domain, requiring transparency, human oversight, and AI-literacy training for staff and students.

Towards inclusive online learning

A responsible approach to GenAI in tertiary education requires institutions to embed accessibility and equity into the very design of their digital learning environments. This begins with applying Universal Design for Learning (UDL) principles, ensuring that courses are structured to provide multiple ways of engaging with content, representing knowledge, and demonstrating understanding. Within this framework, AI tools such as translation services, text simplification, and automated captioning should be made available as opt-in supports, empowering students to tailor their learning experience without imposing unnecessary automation. Equally important is the development of AI literacy among both students and staff, equipping them with the skills to critically evaluate AI outputs , use them ethically, and remain aware of their limitations. Finally, universities must monitor the equity impacts of AI adoption, making sure that technological innovations do not deepen existing divides, and that low-tech alternatives remain accessible for learners with limited connectivity, devices, or willingness to use AI.

Conclusion

Generative AI can be a force multiplier for inclusion in higher education. By providing multilingual access, personalised learning support, and accessible formats, AI has the potential to break down barriers that many students face in online learning. Yet this potential can only be realised if universities place accessibility, equity, and ethics at the core of their digital strategies.

Projects like SOULSS play a unique role: they bring together institutions across Europe to share practices, test inclusive AI solutions, and ensure that the digital future of higher education is both innovative and fair.