A proof of concept codepen by Sarah demonstrating how to use the Azure Computer Vision API to generate alt text for images.
It's as simple as passing the image to the API and returning information about the image including a sentence describing what it thinks is in the image.
Ok it's not perfect, but better than nothing and some of the results are pretty interesting.
https://twitter.com/FrontendNE/status/930120267992616960
https://twitter.com/SniCy/status/930189562923634693
