SightCity 2026 ยท Project preview
Making accessible STEM education easier.
InkluLearn.AI is a planned research and transfer project. It aims to help educators turn visual STEM content into formats that blind and visually impaired learners can use more effectively.
The problem: STEM is often highly visual
Important information in STEM is often presented as diagrams, circuit schematics, curves, tables, formulas, or models. A screen reader cannot automatically understand the subject meaning behind an image.
Why ordinary alt text is often not enough
A short image description may say that a diagram exists. It often does not explain which parts belong together, which signal path matters, which values are important, or what the subject-specific message is.
The idea: understand meaning first, then generate output
InkluLearn.AI follows a semantic-first approach. AI should not only describe the visible surface. It should identify the underlying subject structure and use it to generate accessible alternatives.
How the process could work
- 1
Educators upload existing material such as slides, PDFs, or diagrams.
- 2
AI analyzes the material and identifies elements, relationships, and subject terms.
- 3
A semantic knowledge base is created as a shared source of truth.
- 4
Text, audio explanations, tactile templates, or learning-platform content are generated from that source.
- 5
Humans review and improve the results before they are used in teaching.
Benefits for different audiences
For blind and visually impaired learners
Complex STEM content becomes more accessible and easier to explore independently.
For educators
The effort for accessible alternatives can be reduced without giving up subject quality.
For schools, universities, and training centers
Accessibility can become a more systematic and sustainable part of teaching workflows.
For public bodies and funders
The project connects digital participation, educational equity, and AI transfer in a practical way.
Practical examples
- A circuit diagram is explained through components, nodes, and signal paths.
- A plot is described with axes, trends, peaks, and important points.
- A UML diagram is structured as classes, attributes, methods, and relationships.
- A table or matrix is presented in a screen-reader-friendly order.
Why this consortium is well prepared
THM's Institute of Technique and Informatics contributes expertise in AI, computer vision, modeling, and assistive systems. THM's BliZ contributes long-standing expertise in digital accessibility, inclusive higher education, and tactile and auditory preparation of materials.
How institutions can get involved
We welcome exchange with schools, universities, training centers, associations, public bodies, counseling services, funders, and practice partners. Feedback on real materials, pilot contexts, user tests, and transfer opportunities is especially valuable.
Listen as a podcast
The podcast explains the project idea as accessible audio. The language follows the selected website language when available.
Project Presentation IncluLearn-AI
AI-generated audio summary presenting IncluLearn-AI, its goals, intended context of use, and its approach to accessible learning support.
- File size
- 5.69MB
- Available languages
- German, English, Spanish