Back when generative image AIs first came into general awareness with Dalle-3 and Stable Diffusion and so forth, there was a fad of changing image alt text into deliberately misleading nonsense. Which, of course, was not great for the visually impaired back then either. It was ineffective and short lived, but it's still pretty annoying seeing people proudly declaring their intent to make things deliberately difficult for people who rely on technology for this kind of stuff. So IMO worth pointing out.