Talk:Social media

Archives

---

1, 2

Discovery feed in Wikipedia app similar to social media apps (incl article segments, vids, …)

This is a wish in the Community Wishlist that some people here may be interested in.
It's not social media but the envisioned feature could have some similarities to it, especially if videos relating to Wikipedia (not about random articles but eg short summaries of Wikimania talks) are included and if it's designed so that one can similarly recreationally browse it during commute etc.
Xikipedia and the Recommended Reading List feature of the Wikipedia app are relevant recent developments.

Community Wishlist/W506 (voting open).
Prototyperspective (talk) 12:34, 6 March 2026 (UTC)

AI image use and licencing

Someone posted today on the Wikipedia Reddit about an official WMF social media post which included a fake public domain licence. The way we licence content is a key part of how Wikipedia functions, and various projects take very great care to ensure they are handled correctly, both as a matter of legality and a matter of courtesy to content contributors (those in the movement and those that create freely licenced content elsewhere). The reddit post plausibly suggests the fake licence is the result of AI image use, rather than being deliberate. Working on this assumption, this is another indication that the practical risks of AI image generation. (This is putting aside any philosophical questions on the matter.) I'm not sure how and why it is entering the workflow for social media posts, especially in this case where its use brought no positive benefit to the post. It would perhaps be best to remove it from the workflow to avoid this sort of error. CMD (talk) 15:37, 26 March 2026 (UTC)

Hi @Chipmunkdavis, thanks for reaching out. An upscaler was used to boost the resolution of the image so it would fit an ideal social media size. That's why the bottom right corner of the image looks weird. This wasn't done intentionally as it isn't a tool we normally use. The wrong license information was a mistake – the following image in the carousel was indeed in the public domain, and we ended up using the same information for both. We understand the importance of licensing and recognize what it represents so we decided to delete the post. Please note we're updating our quality control process to avoid mishaps like this in the future. Thank you! LPasqual (WMF) (talk) 14:32, 27 March 2026 (UTC)

I don't really understand how someone can use a new tool unintentionally, but I'm glad it was noticed by the team. CMD (talk) 03:14, 30 March 2026 (UTC)

Facebook Algorithm vs. Identity/Mission

Large pages, in the millions come with their own challenges.

With 5.4M followers and a strategy refresh underway, I wanted to raise two related questions that I think are directly connected.

A) Wikipedia faces a tension most organizations don't: the content formats and emotional triggers that Facebook's algorithm rewards are fundamentally at odds with encyclopedic neutrality. How does the strategy address meaningful engagement without compromising editorial identity?

B) The answer may depend on why people actually follow Wikipedia on Facebook in the first place — mission affinity, content curiosity, habit? A survey or surveys to your @followers could be a fun and powerful strategic input, on FB in particular. Has that research been done, or is it part of the refresh conversation?

Drew-Kelly (talk) 11:49, 25 May 2026 (UTC)