AI 4 Accessibility
Stay informed about how developments in artificial intelligence contribute to accessibility at ETH
AI 4 Accessibility @ ETH
At UTL, we are also pursuing various approaches to harness the possibilities of modern AI to improve accessibility.
In particular, the swissuniversities project "AI4Access" in collaboration with the ETH AI Center with the following focal points should be mentioned here:
- Multimodal representation of non-text content
- Conversion of LaTeX documents (STEM subjects) into accessible formats
- Automatic conversion of handwritten blackboard notes into accessible formats
- Conversion of handwritten additions to presentation slides into accessible formats
- Automatic generation of context-related image descriptions
We are happy to keep you up to date ...
General remarks
The rapid development in the field of artificial intelligence also offers enormous opportunities in the area of digital accessibility.
We are already benefiting from AI in the areas of
- Transcription and translation of audio content
- Automatic generation of image descriptions
- Translations into other languages or into simpler and more easily understandable languages
A word of caution
So far, however, there are no implementations or systems that would render the basic requirements for digital accessibility, as discussed on these pages, obsolete.
Even automated transcriptions and image descriptions always require careful checking by people with the relevant specialist knowledge.
Only the authors of learning materials themselves know what information they want to convey with a particular illustration or graphic, for example.