AI3_alttext

AI3_alttext takes care of creating so-called alt texts, i.e., descriptions of images for people who are blind or visually impaired (and search engines). The complete labeling of all images is a core requirement of the Barrier-Free Information Technology Ordinance, which, due to the Accessibility Enhancement Act, must now be complied with not only by public authorities and institutions providing public services, but also by larger companies. 

AI3_alttext ensures that meeting this requirement does not become a tedious, ongoing task. Even smaller, more energy-efficient, and therefore more cost-effective AI models can generate 80% complaint-free image descriptions with optimized prompts. The remaining 20% are misjudgments that even humans make or that result from stereotypes, which we can avoid with appropriate configuration options (e.g., ignoring gender, skin color, or age). Our goal is to achieve a level of quality that allows automated publication without visual inspection. AI3_alttext can take into account information from captions, such as those describing people or places depicted in the image, during creation. The use of information from other sources, such as team pages, is also conceivable for this purpose.

Other extensions also offer alt text generation, but AI3_alttext allows this task to be largely automated and can run in the background, so that you can move on to the next task instead of waiting for an image description.