And automatically-generated Depicts statements is something WMF already tried and undeployed after significant negative community feedback. Trying something like that again is something that will require significant engagement and development over a long term, and is and should be out of scope for GSoC.
AntiCompositeNumber (they/them) On Wed, Feb 4, 2026, at 04:46, Thiemo Kreuz via Wikitech-l wrote: >> This message (imo) makes it clear that the main problem with this task in >> particular was the presence of agentic AI. > > I'm sorry, but that's besides the point. This is not about what tools > people use. This is about what a project claims to do. The problem > with the project in question > (https://phabricator.wikimedia.org/T414881) is that it aims to change > the way people edit Wikimedia Commons, and the scale at which it aims > to do that. > >> These are/were classification tasks that, for the most part, would help >> community patrollers […] There was no generative AI involved > > I'm afraid the project is deliberately vague about this. Personally I > think this is part of the reason why it was declined. Is it about > patrolling? No, probably not. It specifically says it wants to > "improve the quality of the metadata of the uploaded pictures" > directly during the upload process, as far as I understand it. It's > suggested to generate "depictions" as well as "translations". This is > not merely "classification". This is generating content. > > I hope this helps. > > Thiemo > _______________________________________________ > Wikitech-l mailing list -- [email protected] > To unsubscribe send an email to [email protected] > https://lists.wikimedia.org/postorius/lists/wikitech-l.lists.wikimedia.org/ _______________________________________________ Wikitech-l mailing list -- [email protected] To unsubscribe send an email to [email protected] https://lists.wikimedia.org/postorius/lists/wikitech-l.lists.wikimedia.org/
