Good question.. I’ve wondered about this myself. It appears that the filename is used to create a description for the image. Then after uploading, you can tweak it in the markdown. When viewing a post with an image, you can mouse over it or click on it to look at the image bigger, and it includes the description at bottom. But I don’t know how that description works in screen readers.
If this can work, then for communities where this is especially important you can develop a culture around improving these descriptions, and moderators can also keep an eye out and edit posts to improve descriptions or send PMs to educate users.

@codinghorror is there a place to make specific requests for technical improvements here? Requiring users to rename files or supply markup is simply a non-starter. Those techniques are completely non-discoverable, but even if they weren’t, are also a bigger hurdle than most users will climb. Adding alt text need to be obvious, simple, and (critically) part of the normal process for uploading images, so that users who rely on assistive technologies can participate. IMO the best solution would be to have a pop-up on image upload that asks for descriptive text to be supplied. And at least for our forum I would even like a configuration to make supplying non-empty text mandatory.
@codinghorror I meant to convey that accessibility needs to be a first-class concern that is built directly in to Discourse, not bolted on as an afterthought. (Apart from that, after a recent very bad experience with a non-official plugin and a botched upgrade, we only install official plugins).
Could you point me to a theme component that has these elements? I’d like to play around with one so that I can perhaps sketch out what an alt text / image description theme component workflow would be like.
I do think a plug-in like how events works would be best though, where you insert an image and a form opens up where you type in the alt text and the image description. Then when you hit submit the image is automatically properly formatted.
I suggest starting with a theme component here prior to considering any core changes.
Sorry for the delayed reply. I have always regarded Discourse development as both thoughtful and thorough, and my perception from other threads as also led me to believe that accessibility is a priority.
Which is why I am frankly struggling to understand the reluctance here. I would describe the current situation for users as “nearly impossible to add, even if they want to / remember to” and the situation for admins as “completely impossible to remind / encourage”. Making alt-tags simpler and easier seems like incredibly low-hanging fruit that would benefit accessibility on every Discourse instance. It should be built-in, everywhere. $0.02
You could certainly build a blocking interstitial which prevents posting.. I think you could do this with watched words – since we know the default image description is equal to the filename?
Prompt for descriptions on image uploads (default ON)
Attempt to auto-fill image descriptions using AI service (default OFF)
I would not personally want automated AU labels submitted without human review.
All that said, after three years now, I am somewhat skeptical that Discourse is going to build in functionality to streamline accessibility options for image uploads.
그것도 의미 있는 시도이지만, AI 기능이 없는 포럼에서도 저자를 위한 프롬프트를 제공했으면 합니다.
(디스커스 AI를 탓하려는 것은 아니지만, 지금까지 AI 생성 대안 텍스트는 맥락을 무시하고 매우 평이하거나 부정확한 내용을 제시하는 경우가 많아 개인적으로 선호하지 않습니다. AI는 대량의 과거 콘텐츠에 대한 일괄 처리 같은 용도에는 적합할 수 있습니다. 하지만 글이나 대화에서는 단순한 시각적 설명이 아니라, 이미지가 전달하려는 서사적 의도를 저자가 직접 작성한 대안 텍스트를 선호합니다.)
alt 속성은 이미지 설명으로, 이미지 마크다운 구문 안에 추가하거나 게시글 미리보기의 이미지 섬네일에서 직접 추가할 수 있습니다
마크다운 예시:

저는 이 세부 사항에 대한 제 생각을 계속 다듬어 왔습니다. 시력 있는 사용자에게는 다소 은폐된 alt 텍스트에는 최소한 시각적 묘사가 포함되어야 하며, 서사적 맥락은 이상적으로 figcaption을 사용하여 인라인으로 표현해야 합니다. figcaption은 Discourse에서 아직(아마도?) 사용 가능한 기능이 아닙니다: Support image captions - in addition to alt text.
하지만 현재 우리가 가진 것은 실제로 alt 텍스트인 "이미지 설명"뿐이므로, 이 부분에 대한 프롬프트를 제공하는 것이 여전히 좋을 것입니다.