This is an interesting article on the challenges and opportunities for AI in accessibility. I believe that a major weak point right now is that AI cannot consistently provide meaningful alt text. It struggles to determine what is decorative and meaningful. Also even if it can make that determination it has a hard time determining the author’s intent. I think that humans will need to provide context to the AI. The author’s prompt will help the AI analyze a series of images and provide meaningful alt text in the future. I don’t believe this process will be fully automated for a very long time.