User Moin reports persistent issues with AI captions in Discourse, despite disabling the feature in preferences and selecting ‘Don’t ask again’. The captions keep re-enabling or prompting for permission, even after logout and login. sam acknowledges this as a bug/feature request, noting plans to improve AI captions. keegan investigates, suspecting browser extensions or localStorage issues. Moin confirms the problem occurs across devices and browsers, suggesting Discourse might clear localStorage on logout. Moin later observes that AI captions seem to be disabled, possibly due to a fix or feature removal. sam confirms the feature has been pulled temporarily for re-evaluation, acknowledging both its search benefits and user experience drawbacks, with plans to enhance metadata without impacting users, especially considering visually impaired users.
However, I keep being asked again whether I want to enable captions - whenever I log in and create a post containing an image. So I have to click “Don’t ask again” after every login. That can be several times a day. I expected “don’t ask again” to last a little longer than that. It’s a bit like someone trying to sell you the same thing every time you meet.
Sometimes it’s even worse; captions seem to be enabled without asking me at all. I am not sure how that happens, but they were enabled more than once over the past few weeks - even though I always say no whenever I see the modal.
Can we teach AI-captions to accept that a no is a no?
I think this is unlikely because it happens in all browsers and on both my tablet and laptop. Even with browser apps that I don’t normally use on the tablet.
Is it possible that Discourse deletes discourse_ai-auto-caption-seen on logout?
I was about to report that it had stopped. But I think that didn’t happen because of a fix, but because AI captions are disabled? At least there is no longer a setting in my preferences.
I paid a little more attention to localStorage and it feels like Discourse deletes most entries when I log out. I lose frequently used chat reactions, the state of collapsed and opened sidebar sections, the fact that I enabled push notifications, …
And it makes sense. Because someone else using the same browser might not want to receive push notifications and might prefer different reactions. They may also not have been asked if they want AI captions to be enabled. But it’s still quite annoying to start from square one every day.
I think this is resolved, in that we pulled the feature for now.
Automatic captioning needs a rethink and we needed a breather to rethink the feature.
It is a wonderful feature for search, it is great to be able to type with:images cat and find all the cat pictures.
However the collateral damage here of causing confusing and over long captions is not worth it. We are thinking of ways to enhance the metadata (and handle history) without impacting the user experience.
Additionally, this “extra caption” can be very useful for visually impaired users, so when we revisit we need to take special care with screen readers.
