Contextual Labels

[EstrogenEmpress]

Is your feature request related to a problem? Please describe.

Users are unable to have a ready explanation as to why labels have been applied to certain accounts, mostly by community moderation labelers, but also at times, from Bluesky's own mod service.

Describe the solution you'd like

Additional Ozone and Bluesky native UI support for labeler owners to attribute public-facing "Context/Comment" notes to support moderation calls. On the Ozone side of things, add in feature for labelers to attach comments for public view when applying a label, thereby creating a clear and concise public-facing transparency feature to mod decisions (see Figure 1 - perhaps a check box can be included to "Publish Reason for Action" which would accompany the label to that specific post or account).

This should increase trust and transparency in the composable moderation structure. On the Bluesky UI side, add in a section when a Labeler Comment has been added, below the name attribution (see Figure 2 below, comment could be below the "This label was applied by [labeler service]" section and provide the Reason for Action stated in Ozone if published).

Describe alternatives you've considered

Some have suggested an offsite registry for moderator actions on a per-account basis (i.e. spreadsheet with account names, labels applied, and individual comments from labeler operators). While this may work for certain use cases, speaking as the lead operator for @aegis.blue, we have broken 30,000 individual report activities in just two months - in other words, it adds more work for labeler operators while also becoming exponentially more difficult to maintain with time and labeler popularity, whereas the Comment option integrates this into the labeling flow that already exists.

We also believe that, at least as far as our ethos behind running our labeler is concerned, maintaining such a repository would effectively serve as a catalyst for continued re-litigation and re-ignition of online conflict, which is something we do not want to contribute to.

Additional context

Essentially, we can think of this system like a proto-Community Notes feature. Let's say, for instance, a post promoting a hoax cure for a virus is reported to a Medical Misinformation labeler. In its current state, the labeler can only apply a Misinformation label, then move on. With this feature, the labeler can also attach, for instance, a link to an article or research paper debunking the post, allowing the labeler's users to be not only more informed about the subject matter, but also the labeler's report rationale and increase transparency between users and labelers. This is a trust-building measure first, at least as I'm envisioning it currently.

Fig. 1 Ozone Label UX

Fig. 2 Current Label UX

[Bossett]

I think this should extend one more step: this context should be readily apparent on the appeal screen.

[CallistaAI]

Would love to see this.