Commons talk:Project scope

Proposal: clarify relationship between COM:INUSE and actual practice

Hello all,

I would like to propose a clarification to COM:INUSE, not primarily to change its intent, but to ensure that the written policy accurately reflects current practice.

A recent discussions on Commons:Administrators'_noticeboard/User_problems/Archive_130#COM:INUSE_not_a_suicide_pact? refering to User_talk:Pi.1415926535#File:Unconscience_learning.png_2 demonstrate that, in practice, files that are in use on sister projects are still being deleted based on qualitative judgments (e.g. being considered “AI-generated nonsense”, “decorative”, or lacking meaningful educational value).

This appears to go beyond the current wording of COM:INUSE, which states that:

• files in use are considered in scope,
• even if they are of poor quality or appear to lack educational value,
• and that Commons does not override sister projects’ editorial judgment in such cases.

However, the actual application introduces an additional, implicit criterion: that the use must be considered legitimate or meaningful according to Commons standards, not just the project where the file is used.

This creates a discrepancy between:

• policy (strong deference to in-use status), and
• practice (independent Commons-side evaluation of that use).

I propose that COM:INUSE be clarified to explicitly reflect this, for example by adding wording such as:

Files that are in use on Wikimedia projects are generally considered within scope. However, such use may still be evaluated for legitimacy. Uses that are deemed purely decorative, non-meaningful, or otherwise not contributing to educational understanding may not qualify as valid “in use” for the purposes of this policy.

This is not about arguing for or against specific deletions or specific types of content.

It is about:

• transparency,
• consistency,
• and maintaining trust between Commons and sister projects.

At the moment, contributors—especially newer ones—are guided by a policy that suggests one standard, while encountering a different standard in practice.

This has already led to concrete consequences, such as:

• files being moved to local hosting on Wikibooks instead of Commons,
• adjustments to local guidance (see b:nl:MediaWiki:Uploadtext),
• and growing uncertainty about whether Commons will retain files that are actively used.

If Commons intends to apply a stricter interpretation of “in use”, that is entirely reasonable—but it should be clearly documented.

Otherwise, we risk creating a situation where: Commons appears to defer to sister projects in theory, while overriding them in practice.

Clarifying this would help align expectations and preserve the integrity of cross-project collaboration.

Kind regards, BeeBringer (talk) 13:49, 31 March 2026 (UTC)

Have a consistent robust policy is important. Deleting 3 or a handful of arguably low-quality files in the subjective opinion of some contributors among the over 130 million files on Commons is not important.

Policy should be applied consistently and things can be removed from use in the other projects where Commons does not editorialize, especially not without involving contributors to that project. INUSE doesn't mean bad files must be kept; it means some things need to be decided with the inclusion of or on other projects. Prototyperspective (talk) 14:09, 31 March 2026 (UTC)

Thank you, but I do not think this can be characterized as merely “3 or a handful of files” in any meaningful structural sense.

There is an entire cross-wiki toolchain built around this exact situation. CommonsDelinker exists specifically to remove usages of files from Commons and other Wikimedia wikis after those files are deleted on Commons, so that pages on sister projects do not remain visibly broken. Meta describes this as a service brought to local wikis, and explicitly says that “file delinking” is the remedial action performed after a Commons file has been deleted due to policy infractions. It also maintains a searchable report of removals and renames.

Commons itself describes CommonsDelinker as a bot that “delinks deleted images from Commons and other Wikimedia wikis”. On Meta’s small wiki toolkit page, CommonsDelinker is listed as removing links to files deleted at Commons, with a very large activity figure attached to it.

So my point is not that every deletion of an in-use file is common, nor that policy is routinely ignored in every case. My point is that cross-wiki removal of deleted Commons files is sufficiently routine and structurally expected that Wikimedia has built dedicated infrastructure for it. That makes it difficult to present these cases as if they are negligible edge cases with no practical significance.

And this matters here because once Commons deletes such a file, the consequences are not theoretical:

• the file is automatically removed from pages on sister projects;
• local editors then have to notice that removal, decide whether to re-upload locally, replace it, or leave the page without it;
• and contributors on those projects may never have been meaningfully involved in the Commons-side judgment that triggered the removal in the first place.

That is exactly why clarity is needed.

If the actual practice is that Commons may sometimes override in use status based on an independent qualitative judgment, then that should be stated clearly. If instead the intention is that this should remain exceptional and coordinated with the affected projects, that should also be stated clearly.

At the moment, the existence and function of CommonsDelinker show that the downstream effect on sister projects is real, systematic, and built into the workflow—not something too trivial to document transparently.

Kind regards,

BeeBringer (talk) 14:41, 31 March 2026 (UTC) BeeBringer (talk) 14:41, 31 March 2026 (UTC)

Besides that, I would like to note that your characterization of this as “a handful of files” seems difficult to reconcile with prior discussions on your own talk page (User talk:Prototyperspective).

From a brief review, there are multiple cases where files were explicitly kept due to COM:INUSE, even when concerns about quality or educational value were raised. For example:

• Commons:Deletion requests/File:Rendlesham Forest UFO incident digital CCBY artwork.jpg
• Commons:Deletion requests/File:Art of cow getting abducted by an UFO.jpg

Given these cases, I find it difficult to understand the argument that such situations are negligible, as this appears to be something you have yourself encountered and engaged with in practice and that is only one user of so many.

This suggests that COM:INUSE has been overruled as a meaningful and recurring factor in deletion discussions, not merely as a rare or negligible exception.

For that reason, I believe the question of consistency remains relevant and worth clarifying.

Kind regards,

BeeBringer (talk) 15:09, 31 March 2026 (UTC)

Regarding CommonsDelinker: the uses could be on sandbox pages, talk pages, and user pages. That concerns of subjective opinion about quality were raised doesn't mean the policy shouldn't be adhered to. the question of consistency remains relevant and worth clarifying Policy application should be consistent. Only this way can it be safeguarded and stay robust through the future and not be subject to arbitrary exceptions when some people dislike this or that etc. Prototyperspective (talk) 15:13, 31 March 2026 (UTC)

Thank you — I think we actually agree on the core point.

I fully agree that policy should be applied consistently and not be subject to arbitrary or subjective exceptions.

My concern is precisely that this does not appear to be how COM:INUSE is currently applied in practice. As seen in multiple deletion discussions, administrators and participants sometimes override “in use” status based on qualitative judgments (e.g. “decorative”, “AI slop”, etc.), while in other cases INUSE is treated as decisive.

That creates uncertainty for contributors, because:

• the policy reads as a strong rule,
• but the outcomes do not always reflect that consistently.

So the question becomes:

• should COM:INUSE be applied strictly in all cases, as you suggest, or
• is the current practice (where it can be overridden) the intended interpretation?

In either case, I think we would benefit from making that explicit.

If strict application is the goal, then deviations in deletion discussions may need to be addressed and corrected by administrators.

If not, then the policy text should reflect the actual practice more clearly.

Kind regards,

BeeBringer (talk) 15:30, 31 March 2026 (UTC)

The primary use of Commons:Delinker is when someone uploads a copyright infringing work to Commons, which can't be kept on Commons or many of the projects, which don't allow fair use. We need to stop deleting files that are in use.--Prosfilaes (talk) 03:20, 1 April 2026 (UTC)

• A few responses...
  •  Oppose changing Commons policy in this area. The longstanding arrangement where Commons defers to projects on scope, clearly summarized on COM:SCOPE as It should be stressed that Commons does not overrule other projects about what is in scope is sound, it worked for content before AI and it continues to work now. Many people on Commons seem to have personal preferences around AI images, either strongly opposing or strongly supporting them, and have a desire to enforce their views across other projects. It has never been Commons' place to dictate scope over other projects' usage, and I have not yet seen a strong justification for changing this. -Consigned (talk) 16:49, 18 April 2026 (UTC)
  • I do agree that this is a big problem. The policy is clear: It should be stressed that Commons does not overrule other projects about what is in scope (COM:INUSE). It is not appropriate for practice to stray from policy in this manner - note the {{Policy}} banner at the top of the page: [This page] has wide acceptance among editors and is considered a standard that everyone must follow. If community norms have changed, supporters of the changes should go through the process to update the policy. -Consigned (talk) 12:12, 18 April 2026 (UTC)
  • I'm also very concerned about the adjustments to local guidance that have already happened at b:nl:MediaWiki:Uploadtext, linked in the OP. My browser translates the text in question as Please note: not all files that are considered usable or illustrative on a sister project automatically accepted on Commons. In some cases, for example, with conceptual or context-specific illustrations, it may be more appropriate to upload files locally on Wikibooks so that Commons does not overrule the integrity of the book. This of course contradicts Commons policy and goes against one of the major purposes of Commons - to host free, useful content for other projects. It's hard to believe that practice has strayed so far from policy that other projects are updating their documentation to align with practice rather than policy. This practice being so common does not reflect well on Commons' rigor around policy and respect for the consensus approach that goes into building it. -Consigned (talk) 12:29, 18 April 2026 (UTC)

Concluding remarks

Thank you all for the thoughtful contributions to this discussion, both here and on the Administrators' noticeboard.

It is clear from this thread that there is a strong consensus that COM:INUSE remains a vital policy for the integrity of the Wikimedia ecosystem. As noted, the policy explicitly states: It should be stressed that Commons does not overrule other projects about what is in scope.

However, I started this proposal because of the growing gap between that written policy and the reality of enforcement. When an administrator explicitly states that they view following this policy as a "suicide pact" and feels empowered to delete "in-use" files based on subjective labels like "garbage" or "nonsense," the trust between projects breaks down.

For a contributor on a sister project, the situation is now as follows:

• The Policy says: "Your file is safe because it is in use and provides educational value to a sister project."
• The Practice is: "An administrator may delete it anyway if they personally dislike the content, forcing you into a complex and time-consuming undeletion process."

While I still believe the policy text could benefit from clarification to prevent this "arbitrary enforcement," I have noted the arguments that the policy is already "clear enough." The core problem, it seems, is not the text itself, but the occasional willingness to bypass it.

As a result of this unpredictability, I—and others—feel forced to prioritize the integrity of our local projects (such as Wikibooks) by hosting files locally (see b:nl:MediaWiki:Uploadtext). It is a loss for the "central repository" mission of Commons, but a necessary step when the established rules are treated as optional.

I will not be pursuing a formal change at this moment, as this discussion has sufficiently highlighted that the issue lies in enforcement rather than the spirit of the rule. I want to thank everyone who stood up for the principle of cross-project respect. I hope that, in the future, the "wide acceptance" of this policy will be consistently reflected in the actions of all administrators.

Kind regards, BeeBringer (talk) 20:27, 21 April 2026 (UTC)

I would characterize the current practice as, if keeping a file causes significant harm, most frequently copyright violation, to the extent that outweighs the importance of not overriding sister projects' editorial decisions, an administrator may delete an in-use file. What constitutes significant harm can be more inconsistent between administrators than we hope, and that is where elaborations would help, though. whym (talk) 11:19, 30 April 2026 (UTC)

I appreciate your perspective, whym. If the practice were indeed limited to cases of 'significant harm' like copyright violations, there wouldn't be much of a conflict.

Unfortunately, recent discussions—such as the one at Commons:Administrators'_noticeboard/User_problems/Archive_130#COM:INUSE_not_a_suicide_pact?—show a different reality. In practice, many files are being deleted based on subjective labels like 'AI slop' or 'decorative,' even when they are actively providing educational value elsewhere.

This unpredictability has real-world consequences. Personally, I have reached the point where I will no longer host AI-generated images for Wikibooks on Commons. Even if the Wikibooks community finds the images valuable, and they cause no real harm according to the Commons official standards, I still cannot be certain that Commons will respect that 'in-use' status. This forced fragmentation—operating on multiple levels just to protect the integrity of a book—is exactly why the Dutch Wikibooks now advises local uploads instead of using Commons (b:nl:MediaWiki:Uploadtext).

When Commons overrides the editorial judgment of a sister project in this way, it stops being a central repository and instead becomes a source of instability for the very projects it is meant to support." BeeBringer (talk) 05:27, 1 May 2026 (UTC)

To seek definitive clarity on this matter, I have followed the suggestion of the deleting administrator and opened a formal undeletion request at Commons:Undeletion_requests/Current_requests#File:Unconscience_learning.png.

To be clear: I have already improved the file locally on Wikibooks and clarified its symbolic meaning there, so from a purely practical standpoint for my project, the undeletion is not strictly necessary. However, I am pursuing this as a test case to see if COM:INUSE is still a functioning policy or if it has indeed become a "dead letter" that can be ignored at administrative discretion.

If the community or administrators decide that subjective labels like "garbage" or "nonsense" can override the editorial judgment of a sister project, then we will at least have the transparency I was asking for: that Commons no longer guarantees stability for files used by its sister projects. I hope this request will show whether the policy is actually enforced. --BeeBringer (talk) 06:02, 1 May 2026 (UTC)

After attempting to follow the official procedures via an Commons:Undeletion_requests/Archive/2026-05#File:Unconscience_learning.png, I have reached a clear conclusion for this discussion.

The reaction to my request was not based on policy, but on personal hostility toward AI used. Instead of a professional debate on COM:INUSE, I was met with:

* Intimidation: Threats of a "referendum" on my behavior for quoting policy.

* Gaslighting: Denial of documented facts in the edit history.

* Isolation: Attempts to force the discussion into a non-English environment to hide it from the broader community.

* Collusion: Pinging a specific circle of medestanders to drown out policy-based arguments.

It is now undeniably clear that a group of administrators is so fundamentally opposed to AI-generated content that they are willing to ignore the editorial judgment of sister projects and the explicit text of COM:INUSE.

Because Commons no longer provides the certainty and protection that its own policies promise, I will no longer waste energy on this project. This discussion serves as a record for anyone wondering why sister projects (like the Dutch Wikibooks) are increasingly forced to bypass Commons to protect their own editorial integrity. The gap between written policy and actual practice has become too wide to bridge. Im am veruy soory about this conclusion because I think commons deserves all the respect. Kind regards, --BeeBringer (talk) 08:13, 1 May 2026 (UTC)

Postscript: Shortly after I posted this conclusion, the undeletion request was summarily closed on procedural grounds after only a few hours—bypassing the standard duration for such discussions. This confirms the final strategy: Bureaucratic Burial. By moving the discussion and labeling it "out of scope," the core issue of COM:INUSE is effectively buried to avoid a substantive debate. This is the final proof that the current system prioritizes administrators convenience over policy compliance. I am now officially leaving this matter. --BeeBringer (talk) 08:53, 1 May 2026 (UTC)

Even if this very long argument, based on a single AI-generated image and taken to an extreme degree, were persuasive, I don't think you're going to get much engagement with apparently AI-generated talk page text. — Rhododendrites ^talk |  12:10, 1 May 2026 (UTC)

References

SPAM and use of commercial photos, sourced from product makers

Where should we draw the line on COM:SPAM? In particular, product photographs from their makers.

See:

File:Kidvelo_Rookie_12_Balance_Bike.png

Commons:Deletion_requests/File:Kidvelo_Rookie_12_Balance_Bike.png

User_talk:Polarlys#Commons:Deletion_requests/File:Kidvelo_Rookie_12_Balance_Bike.png

This is a commercial product, a notable class of products and a one we don't have many clear photos of. This is a commercial branded one, and a quality photo from their maker. Should this be deleted as simple spam? @Polarlys: Andy Dingley (talk) 13:14, 20 April 2026 (UTC)

@Andy Dingley: I think images and other media of commercial products from their manufacturers can be kept if we have VRT permission to that effect, or a verified license on the manufacturer's site or social media. Promotional text can be removed under COM:CSD#G10.   — 🇺🇦Jeff G. ツ please ping or talk to me🇺🇦 13:25, 20 April 2026

AFAIR (it's deleted now, I can't see this) the licensing for this image was evidently valid and was never challenged by any party. It was deleted for being 'undeniable spam'. Obviously I'd already denied this, but that was just the regular @Dronebogus: hyperbole.

The point here though is the general one, so please don't let's get derailed into arguing an irrelevant point about one specific example (where that didn't even apply):

Are we doing deletions on the basis that any commercial product image from a manufacturer is 'undeniable spam' ? Because IMHO, that's an unconstructive choice, and certainly if it's assumed as a blanket reaction. Andy Dingley (talk) 14:56, 20 April 2026 (UTC)

The argument that a photo of a product taken by the manufacturer is out of scope seems like a misapplication of COM:SPAM. If a product is notable, or even of a notable class of products, a company's own photos will often be very good depictions of those products. So in scope. Not that different from when a celebrity uploads a professional photo to use in place of a lower quality free photo -- yes, I guess it's "promotional" but it's also educationally useful. If the subject isn't notable, then it's worth having a conversation, and certainly if the promotional nature of the photo led to manipulations that harm educational usefulness, then that's a separate issue. The photo above is now a redlink, though, so I can't see what we're talking about. — Rhododendrites ^talk |  17:21, 20 April 2026 (UTC)

If a commercial image holds no educational value, I don't see any reason to keep it. At that point, it’s basically just free advertising.

We have plenty of photos of certain products taken by actual users that look natural and fit the context. However, when we see a perfect shot with professional cropping and studio lighting, it feels out of place and appears to have a direct promotional intent. To me, the level of polish suggests the intent is promotional or perhaps a copyright violation by the user, rather than being truly educational. RiadS99 (talk) 21:03, 21 April 2026 (UTC)

@RiadS99: when we see a perfect shot with professional cropping and studio lighting, it feels out of place... That seems very odd to me. I'm all for shots taken "in the wild"—that's what I mostly do—but surely you aren't saying that studio photography is not appropriate to Commons, so what are you saying? Do you have a similar issue with us using official portraits of politicians/headshots of actors/band publicity photos when they are available on a "free" basis? - Jmabel ! talk 06:13, 22 April 2026 (UTC)

@Jmabel Honestly, sometimes (very professional pictures, depending on the context of course) feel to me a bit like a form of free advertising.

Take this photo, for example: clearly, the user was asked to upload this professional picture of the journalist (plus his signature) and then add it to the article. Even though the process here is technically correct: proper license, EXIF data, etc.I still get the feeling that the platform was used purely for self-advertising purposes, rather than following the natural process, because usually, users either find pictures of journalists with free licenses elsewhere and try to add them, or a journalist naturally comes across a photo taken by someone else. RiadS99 (talk) 15:55, 23 April 2026 (UTC)

For the purpose of this discussion, I have restored the image. @Rhododendrites: Regarding the discussion I repeat myself: “what to say about the only and unused upload of an account with the file description *** Bikes are a new brand specialising in the manufacture of balance bikes for children.”?' To me, this is Advertising or self-promotion as outlined by the policy. I have no strong feelings about this, however. --Polarlys (talk) 18:37, 20 April 2026 (UTC)

Thanks for the ping. If I had participated in the DR, I'd probably err on the side of keep. Presumably any photo that just shows a particular kind of bike on a white background could be considered product photography and thus in theory promotional. But it's also ... a useful kind of picture to have. It's not like this company has inundated us with a thousand photos of their bikes -- there are only three other bikes on clean backgrounds in the category, and presumably they could be deleted as promotional, too? — Rhododendrites ^talk |  18:40, 20 April 2026 (UTC)

I added a note about the restoration to the subpage. The file description page does not appear to be overly promotional.   — 🇺🇦Jeff G. ツ please ping or talk to me🇺🇦 20:51, 20 April 2026 (UTC)

I've removed the promotional description. I think the file is fine. I'd love to have a good, free-licensed picture like this for every model of bicycle, and the fact that it came from the manufacturer does not strike me as in any way a problem. Would that more manufacturers would free-license things like this. - Jmabel ! talk 23:18, 20 April 2026 (UTC)

I completely agree. Assuming the license is valid, I think this is a good photo to have in our collection. I would welcome similar photos from other manufacturers of all kinds of products. —Mx. Granger (talk  · contribs) 02:27, 21 April 2026 (UTC)

Generally speaking, I agree. That being said, I've also seen some product photos which were very clearly promotional and which seemed unlikely to have educational value, e.g. photos of bottles of liquor or perfume. Omphalographer (talk) 20:37, 21 April 2026 (UTC)

Bottles of liquor and perfume don't have educational value? --Carnildo (talk) 21:35, 8 May 2026 (UTC)

The images I had in mind weren't simply photos of a bottle. They were photos of e.g. a liquor bottle on a table in a lounge scene, surrounded by images of awards. Placing them in an article would have amounted to running an advertisement for the product. Omphalographer (talk) 22:05, 8 May 2026 (UTC)

Visual depiction of violence

Commons:Village_pump#c-WMFOffice-20260319222200-Office_action:_Removal_of_file

The video in that case is a cctv footage from Texas, USA, that shows a person was beheaded and then the head was rolled on the ground. No one involved was an Australian citizen. That video was taken down by Wikimedia Foundation after a request from the Australian government.

In the discussion, users have different opinions on whether that video would be kept or deleted according to Commons community consensus, so these questions are raised.

Where is the boundary of Commons:Project scope with regard to this specific kind of files, which contain "visual depiction of violence"?

Killing of Nguyễn Xuân Đạt (Q137217543): 2025 killing in Vietnam might be a good litmus test. There are photos and videos of the murder process. (Caution if you google: extremely graphic and unsettling scenes.) It is a notable case, but it is not related to terrorism, or hatred against gender, race or religion... so hate crime/terrorism/sedition laws do not apply.

Is "the boundary" (between acceptable and unacceptable to be available on Commons)

1. before "any blood or injury or corpse can be seen" (so no photos of such things allowed)
2. between photos and videos
3. between edited version and original version of the depiction (like how the BBC often pixelate, blur, muffle or cut out the most graphic part of a video)
4. allowing audio and video in unedited, high-resolution and full-length form?

Is "a person's intention to harm" also a factor? Are files treated differently if they show graphic scenes of pure accidents (avalanche, wildfire, lightning strikes... so no human cause), or animals attacking humans?

Then what about depiction of animal violence deliberately caused by humans e.g. cockfight (Q319771): blood sport of two roosters set on each other dog fighting (Q1066914): dog bloodsport? (No human is harmed but other animals are harmed on purpose.)

Is time also a factor? Does the community treat such files differently if the file was made a long time ago (all persons involved; and all spouses, children, parents and siblings of the persons involved, are dead by now)?

(Obviously, there are plenty of non-visual (e.g. textual) depiction of violence that is on Commons. Many books describe such things in detail. It seems few or no such files have been objected to yet.) RoyZuo (talk) 09:32, 22 April 2026 (UTC)

I don't read that event as something that demands clarification of our scope policy. It sits somewhere between the application of COM:PEOPLE and the fuzzy area of takedown requests and office actions that are always a case-by-case basis. One thing seems clear: the WMF is telling us, in so many words, that any time there's a possible COM:DIGNITY/COM:PEOPLE concern, we really need to articulate a strong rationale for an overwhelming educational value. I don't think you can boil that down to e.g. "blood or no blood" and don't think it makes sense to get into intentionality or specific cases. We could strengthen the language here about COM:PEOPLE trumping COM:SCOPE, but otherwise if there are edits to be made it's to the COM:PEOPLE page IMO. — Rhododendrites ^talk |  17:53, 22 April 2026 (UTC)

Addition

I have added "Project pages created by and edited only by the file's uploader." as their seems to be a consensus for this. .     Jim . . . (Jameslwoodward) (talk to me) 13:17, 7 May 2026 (UTC)

Endorsed (just so that this section isn't created and edited only by James). - Jmabel ! talk 20:19, 7 May 2026 (UTC)

Where was this discussed? -Consigned (talk) 00:37, 9 May 2026 (UTC)

Would also like to know. - Alexis Jazz ^{ping plz} 07:04, 19 May 2026 (UTC)

Discussion at Commons:Village pump#Draft for an INUSE exception addendum

 You are invited to join the discussion at Commons:Village pump#Draft for an INUSE exception addendum. - Alexis Jazz ^{ping plz} 19:57, 18 May 2026 (UTC)