Wikipedia:Administrator elections/May 2026/Candidates/Bobby Cohn

The following discussion is preserved as an archive of a successful administrator election candidacy. Please do not modify it.

Final (364/126/56) (S/A/O); See official results msk 02:20, 20 May 2026 (UTC)

Nomination

Bobby Cohn (talk · contribs) – It gives me great pleasure to present Bobby Cohn as a candidate for adminship. Bobby has been a maintenance workhorse since he began editing in 2023. He work includes lots of careful new page review (he has a lengthy and almost exclusively red CSD log) and work at RM/TR, and along the way he's made a number of reports to AIV and UAA. While doing so he has kept in touch with the content side of the project, creating or improving a few dozen articles on topics as diverse as a book about cybersecurity, a presidential speech on foreign policy, and a sporting subculture. Bobby is a considerate and thoughtful editor, always willing to take constructive criticism. He will make an excellent administrator. Vanamonde93 (talk) 02:45, 5 May 2026 (UTC)

Co-nomination

But wait, there's more! Bobby is also a highly experienced AFC reviewer, often making as many as 200 reviews a month. These aren't quick front-of-queue declines, either; he leaves helpful, substantive comments, or even improves the article himself so that it can be accepted. He responds to questions about these reviews thoroughly, patiently, and kindly (eg , ). This goes far beyond the necessary, a good sign for WP:ADMINCOND, and you'll find similar thoroughness and attention to detail in his AFD !votes and SPI filings. It's easy to get burned out doing this kind of work - and, as you'll read below, Bobby has. His demonstrated ability to take criticism to heart and refocus on what matters makes it clear that he's ready to handle whatever adminship can throw at him. Let's give him the mop. asilvering (talk) 22:05, 5 May 2026 (UTC)

Candidate, please indicate acceptance of the nomination here: I humbly accept the nomination. Bobby Cohn 🍁 (talk) 16:48, 5 May 2026 (UTC)

Please disclose whether you have ever edited Wikipedia for pay: I have never edited Wikipedia for pay.

Questions for the candidate

Dear candidate, thank you for offering to serve Wikipedia as an administrator. Please answer these questions to provide guidance for participants:

1. Why are you interested in becoming an administrator?

A: My interest in becoming an administrator is in a desire to continue to help out the project by reducing the load of other editors in areas where I am already active. Sometimes I find myself adding to the backlog rather than resolving them, most notably at UAA but in other areas of the project including AfC and NPP where advance permissions are sometimes required for cleanup. I would use additional tools to aid in my work that I come across in NPP, and viewing deleted content to assist in anti-vandal or general newcomer guidance work. Additionally, because I have experience in participating in AfDs and reading and summarising clear consensus statements for closes (like at RMs), I could also see myself doing closing work at AfD. I would also benefit from having access to more information in deleted content when filing SPI reports. More broadly, I have tried to reduce the workload in other areas where I currently hold advanced permissions, and I would extend that same philosophy with a larger toolset.

2. What are your best contributions to Wikipedia, and why?

A: I think some of my best contributions to Wikipedia might be the guidance I've given to new editors that is tucked away in talk page archives or in discussions with new editors on their talk page. I often take the time to break down policies and guidelines (be it content, referencing, notability related) to new editors who may have the requisite knowledge or access to source materials that I or other editors may not, but struggle with translating it into a Wikipedia-appropriate article. I view it as a bit of a butterfly or knock-on effect: I can do more good (writing, editing, etc.) through others with this compounding principle.

In a more traditional Wiki-CV sense, I am also really proud of my two GAs; one from scratch and another translated. Personally, I find it very fulfilling to be able to focus completely on an article and its source material and then have that reviewed by another set of eyes. It gives a great sense of satisfaction knowing that together, we've left an article in a state of "polished and ready for the masses".

Another thing that I really take pride in doing is bringing a strong assessment to AfDs wherever possible. I think it's good to have a balance of creation and deletion work, and doing both has instilled a solid philosophy of "sourcing is king" to my work. To that end, I pride myself in giving {{source assess table}} reviews of sourcing in reviews or AfD discussions. I do this where I think there's an argument to be made to try and convince others and change a developing consensus, or to shine a light on something that might be overlooked (here's an example). To that end, I've also recently begun developing a user script to build assessment tables to reduce a barrier for others to contribute as well.

3. Have you been in any conflicts over editing in the past or have other users caused you stress? How have you dealt with it and how will you deal with it in the future?

A: Last year, I had a conflict with a new editor who appeared to be using LLMs to both develop content and communicate with others. My concerns were about (what I viewed as) the developing mess across the mainspace and other content and project areas; and my frustration stemmed from the argumentative communication and stonewalling that were coming from the LLM. When I was concerned about problems continuing in the mainspace, I escalated to ANI out of frustration. Another editor pointed out that, whether or not I was correct, I was still acting bitey towards a newcomer. I decided that I needed to take a step back and I took a month off to focus on re-centering myself and reflecting on what I wanted to get out of editing.

Since coming back, I've adjusted how I approach communication with new editors, regardless of what might be the cause of frustration, and I'm satisfied that I've been able to deliver clear-headed and constructive advice in the face of situations that may be a source of conflict. Looking back, I definitely regret last year's actions, but I use that feeling as guidance when I'm dealing with potentially stressful situations, and try and hold myself to the higher standard of future me.

---

You may ask optional questions below. There is a limit of two questions per editor. Multi-part questions are disallowed, but you are allowed to ask follow-up questions related to previous questions. Make sure to use level-four section headers, not boldface.

Optional questions from Robert McClenon

4. This question is about the use of artificial intelligence in Wikipedia in both articles and discussions. What do you as an administrator plan to do to limit the use of artificial intelligence?

A: That's an interesting question from a technical perspective. Aside from enforcing the current policies on AI usage as written, I don't believe there's much that administrators can do to systematically limit the use of AI on Wikipedia. I will say that I believe the edit filter managers have done excellent work identifying and tagging AI-generated content where reasonably possible, though I suspect that as the technology continues to develop, it will be an arms race to try and identify new telltale signs of AI.

I think the more important role for administrators is ensuring that AI does not improperly influence the decision-making process that is seen on a day-to-day basis, like closing discussions and determining consensus. This is similar to work I already encounter at AfC and NPP. When looking at new articles, it is important to closely review anything that may come from an AI hallucination (this is part of the problem with AI, it can be generated a lot faster and at a bigger scale than it can be reviewed by volunteers). Administrators today have the difficult task of ensuring that AI-generated arguments, which may sound plausible but are technically incorrect or are a misrepresentation of cited policies and guidelines, do not influence their decision making improperly.

If I were entrusted with the tools, I would make sure to enforce the policies as the community currently expects administrators to do. I believe this includes continuing to stay aware of advancing technologies and changes to the telltale signs of AI writing and to diligently but responsibly make sure that unreviewed or incorrect writing is not kept on the project and that it is not used to influence the decisions of myself or others. It also involves staying informed of developing policies as they are reformed and adapted by the community.

5. This question is about unregistered editors and the temporary accounts assigned to unregistered editors. In closing discussions such as AFDs, RFCs, and ANI threads, should the arguments of unregistered editors be given the same weight as registered editors, or less weight, or should they be ignored?

A: On their face, all arguments should be weighted based on the merits, regardless of the type of account (or edit count of a registered user, for that matter) making that argument. What is important in determining consensus is the policies and guidelines that are cited in that argument and the sources present both in the argument and article, and determining how well both the P&Gs and sources may or may not support the given argument. Unregistered users play a valuable role in helping the project out, and I will always try to help a user complete a task for which they may not have the requisite tools, that is nonetheless productive (like, for example, opening an AfD discussion).

Of course, like any rule, there are exceptions. Closers should always have an eye out for the potential for sockpuppetry, and in topic areas subject to editing restrictions where unregistered or non-EC editors are not permitted to participate, those comments are typically struck and closing administrators are instructed to disregard them when assessing consensus.

Optional questions from I2Overcome

6. Over the years, Wikipedia has endured many challenges. What do you think is the greatest challenge Wikipedia faces today, and why?

A: I think the greatest challenge Wikipedia faces today is editor retention. Wikipedia ultimately depends on a relatively small number of volunteers who spend enormous amounts of time maintaining articles and writing new content. Some of this work involves highly technical skills, such as developing and maintaining tools and bots, or writing modules and templates. Many of the other challenges we face (coordinated disruption, AI generated content in both source material and article writing, declining quality of research tools, paywalling of accessible source materials, and the doxxing and targeting of editors) ultimately place additional strain on that volunteer corps and add to the growing backlogs requiring more attention from our editors.

Wikipedia has historically been very good at surviving vandalism and external challenges. What concerns me more is the long-term sustainability of the community itself. If experienced editors become burned out faster than new editors can successfully integrate into the project, it becomes harder to maintain the quality and reliability that readers expect, and we may lose the reputation we've worked hard to earn, that separates us from–say–Grokipedia.

7. If you have not already explained elsewhere, please explain how you would approach this challenge as an administrator.

A: As an administrator, I would approach the editor retention challenge by focusing on reducing unnecessary friction in areas where administrators can meaningfully help. That includes acting professionally and calmly in contentious discussions, helping new editors understand policy rather than simply templating them, and making administrative spaces feel fair, predictable, and transparent.

This is a practice that I continuously try to implement in my work regardless of the permissions that I hold. I often offer guidance to new users that is particular to their specific subject area or draft article. My hope is that it (1) tells the user “they have read my work and are offering me advice specific to improving my writing directly” instead of templates that may have many links but result in banner-blindness, (2) results in improved content, and (3) is more likely to have that editor return to continue to contribute.

I would look forward to bringing this philosophy to additional areas that are routinely handled by administrators. This would include following up on CSD-deleted material, user talk interventions, and noticeboard threads, where the tone set by admins in any given discussion shapes how welcoming or unwelcoming Wikipedia feels to newer contributors, or where direct but specific advice may help in better addressing problematic behaviour. If admins model patience and good faith, that sets a standard that others are more likely to follow.

Optional questions from Left guide

8. When fielding reports at WP:UAA as an admin, how would you decide whether to a) block on sight, b) attempt to warn or discuss with the user, or c) defer the matter for community discussion?

A: All three of your options correlate with the principle that "are used to prevent damage or disruption to Wikipedia, not to punish" and from that, I interpret it as a need to consider urgency and severity when using any admin tool.

Most of my experience at UAA stems from reporting accounts that combine a promotional username (i.e.: a corporate name) with promotional editing about the same, typically whose material will also qualify for G11. Those are the instances I believe blocks are acceptable. The standard warning for this type of block explains the issue of account names and PAID editing requirements as given by the foundation; editors engaged in paid promotion should (ideally) be familiar with these policies before contributing, and can be expected to follow the explanation as given in the block notice. Because a block in this scenario requires both a promotional username and promotional edits, I'm less concerned that a good-faithed editor would be incorrectly caught up in one. Blatantly offensive names that slip through the regex filter where the username and would clearly cause disruption are the other obvious case for an immediate block.

Situations calling for warnings or discussion are often less pressing in nature and not as clear-cut. Corporate names of users who are otherwise contributing appropriate content in accordance with our COI guidelines (i.e.: their material would not qualify for G11) are better handled through talk page outreach first. Ideally, these situations would be able to be addressed by notice and discussions and only if those went unaddressed or ignored, would a particular type of block be necessary (see {{uw-softerblock}}). Similarly, usernames that aren't clearly disruptive and for which a good-faithed argument could be made deserve a discussion on their talk page. Ideally, we would avoid driving away well-meaning editors over a correctable mistake or an unintentional error by someone new who wasn't entirely aware of our username policy.

Deferring the matter for community discussion would be the appropriate action in contested cases where reasonable editors could disagree, but only after first attempting to discuss with the user on their talk page. If an appropriate solution can be achieved by well meaning and calm editors in respectful dialogue, it reduces the bureaucratic headache for everyone involved and lets everyone get back to focusing on building the encyclopedia.

9. Under what circumstances would you consider it appropriate to block a user who has been reported to WP:AIV but hasn't received any warnings?

A: The standard reasoning behind requiring warnings before a block is sound: editors (especially new editors) may not know they are doing anything wrong and our tiered warning gives them a fair chance to self-correct before a more serious consequence. I do not want to abandon that principle lightly. An unwarned block about someone confused about policy, especially one who is otherwise on the path to becoming a productive editor, will cause serious frustration if not the loss of an editor entirely.

However, there may be times when other policies might necessitate taking a more definitive approach to prevent harm to the project or other users. If the nature of the vandalism makes it impossible to assume good faith, such as slurs, targeted harassment, harmful hoaxes into BLPs, then I could envision a scenario where the actions may necessitate an immediate block. But on the whole, I would be reluctant to block without at least attempting contact. I remember reading somewhere that a warning costs very little, and in the unfortunate event that a block is necessitated later, it only bolsters the validity of the subsequent block (see also: WP:ROPE).

Lastly, there is an area of the project where it may be more beneficial to block without warning: clear sockpuppet accounts whose username or vandalism patterns match an existing SPI or LTA. Here it would be more beneficial to block without spilling WP:BEANS. But in areas where discretion may be preferable, I would likely first consult admins who have more tools than I (CUs) to handle and confirm any suspicions, especially if the matter weren't pressing.

Optional questions from George Ho

10. What do you think about ArbCom's recently passed principle "Focusing on content isn't enough"?

A: It looks to me like it is important to read it together with immediately preceding "Being right isn't enough" too. Together, I think the wording fairly represents an intersection between our civility policy and a realization that editors working in collaborative spaces will naturally run into disagreements that require communication or bureaucratic dispute resolution procedures. Inherently, human nature and our ability to communicate in a text-based medium here may also be a limiting factor but ultimately, I think if you are communicating about another editor's work it is important to be pleasant and respectful.

11. What about Christophe Henner's LLM-generated usage of an LLM in his Signpost article?

A: That's an interesting article that I hadn't seen, thank you for linking it. On a personal note, I am a bit of a stats nerd, so I always appreciate digging into the numbers. I am always cautious that numbers are always only part of the story, and can be used to support many (sometimes even conflicting) arguments. I'm not saying I think the author was misrepresenting anything, rather it's something I always consider when I see people try and tell a story with numbers. With regard to that article, I think it raises some interesting questions and my mind went to "oh, there's probably a study design that could test that" or "I bet there's data somewhere that could tease out some confounding or explanatory variables here".

On the whole, I'm glad there are people interested in monitoring these things and who are clearly interested in the long-term health of the project. With regard to LLM usage, I did try to approach the article with fresh eyes, but I got caught up in some of the more tell-tale signs, and I think it will be unfortunate if that ends up being the main takeaway.

Optional question from Significa liberdade

12. Administrators are likely to receive personal attacks, including name-calling and threats of violence. How might you respond to such attacks personally and on-wiki?

A: Thankfully, I've never found these types of situations stressful. I like to try to remember "you don't know what someone else is going through on the other side of the screen" and escalating is always going to be counter productive. On-wiki, it would depend on the source. If it stemmed from a content dispute I'm involved in, I'm plainly too close to it to use admin tools. I'd step back, note my involvement, and leave any action to another administrator. If the situation warranted oversight, I'd flag it for that team discreetly. Beyond that, I'd do no more than I would as a regular editor.

If it isn't coming from a content dispute but rather straightforward abuse or petty provocation (and assuming there isn't an WP:ADMINACCT to address), I'm a big fan of the WP:DENY policy, as engaging rarely serves anyone and just rewards the behaviour. The same recusal logic applies here too: if admin tools are needed, someone uninvolved should wield them.

However, threats of violence are different and should not be casually ignored. Whether to myself or others, I would follow the procedure as outlined on Wikipedia:Responding to threats of harm closely.

Questions from Carrite

13. How and when did you first become involved at Wikipedia?

A: I first became involved on Wikipedia seriously when I noticed a gap in Wikipedia's coverage. I spent probably about a week collecting sources, furiously reading project pages on advice about writing, looking at other article's text in the source editor to figure out how to format and write an article before submitting it to AfC. I loved every moment of it, from learning more about the topic as I researched, to grouping similar ideas and fleshing out article structure and writing. Once I thought I had a handle on it, I wanted to learn more and become more involved to see what else I could help out with, and that has more or less been my guiding principle since.

14. Have you ever edited at Wikipedia under any other account names, and if so, what is the backstory of the multiple accounts, change of name, or clean-start?

A: Back in high school, I had an account that I used to fix a handful of typos. I didn't use it long enough to remember the username. I returned to Wikipedia (as mentioned above) when I was mostly through grad school.

I have done a name change once on this account. I took the name (well, similar) from a foil in a Hemingway novel; a character who probably takes himself too seriously, thinks he is better read, more well-spoken, and tougher than he actually is. I think that it's a funny but subtle self-deprecating reference to carry on a project where we're all just here to build an encyclopedia, but there isn't much more to read into it than that.

Optional question from Spravato

15. If you could change any Wikipedia policy, what would you change?

A: Nothing jumps out at me as working so poorly that I would feel the need to unilaterally decided to change it. Frankly, I would hate the idea of imposing my will on the community, part of the reason why the project works well is that any proposed changes get to be reviewed and discussed by editors with more experiences and varied skills than one editor can bring to the table. No editor, regardless of technical permissions, is more important than another, and it's precisely this process that gives any credibility to our P&Gs.

Optional question from VortexPhantom

16. If a company offers you to create a article on them,and you find that the company meets WP:NCORP, in return of $100. What you will do and why?

A: I would probably get a good chuckle out of it and then immediately decline for multiple reasons, least of which because I'd rather spend any writing time I have on articles that I want to work on.

In all seriousness, I have no intention of ever editing for pay; nor would I use administrator tools for pay if entrusted with them, for that matter. I would make them aware of the Foundation's Terms of Use at WP:PAID and our WP:COI guidelines, and probably caution them on WP:SCAMs. I might even keep an eye on the article title and look out for UPE and PROMO as well.

17. Share your thoughts on this newly passed guideline WP:LLMPROD where a user who has determined to have a history of AI use, a series of their contributions may be assumed to violate Wikipedia's policies and guidelines, because LLM generated texts generally violates Wikipedia's core policies. In this case, edits to existing pages may be reverted without review, and pages where the editor is the only significant contributor may be nominated for a special proposed deletion process, LLMPROD, subject to the certain requirements. Do you think this is fair and effective against increasing AI abuse in Wikipedia?

A: I think it reflects the needs of the project to ensure truth and accuracy through WP:V and the unfortunate reality of the limitations of editor time. I've mentioned in my earlier responses that AI-generated content can be made faster than it can be reviewed, while volunteer time is limited and valuable. In an ideal world where we have infinite resources, every suspected AI-generated addition could be reviewed and verified for truth and if found incorrect, be addressed on a case by case basis. But we know that we have to balance very important considerations, especially in areas like BLP where incorrect information can have serious repercussions. My hope is that it wouldn't be used punitively but in defence of the project and out of an abundance of caution.

Discussion

• Links for Bobby Cohn: Bobby Cohn (talk · contribs · deleted · count · AfD · logs · block log · lu · rfar · spi)
• Edit summary usage for Bobby Cohn

---

Please keep discussion constructive and civil. If you are unfamiliar with the nominee, please thoroughly review his contributions before commenting.

I liked the article on stretcher railings (about fences in England made of repurposed medical stretchers) when it showed up at DYK, and feel like he handled critical feedback really well in a way that resulted in a better article. Looking at some other areas, I see hundreds of UAA reports, loads of AfC work, many messages helping out newer AfC editors, and a diverse range of navigation box cleanup. Rjjⁱⁱⁱ (talk) 00:04, 8 May 2026 (UTC)

AfD record: 77.1% match rate, n of 407 comments on unique AfD pages, with 8 keep !votes to 167 delete !votes. 210 pages included “no discernible vote by this user”. Subjective comment: After looking at a sample of listed pages, many of the “no discernible vote” items may be Bobby nominating pages for deletion. Of these, the sample I looked at were either deleted, redirected, or closed without consensus. Other !votes appeared to have substantial content, rather than merely stacking on !votes. Significa liberdade (she/her) (talk) 21:24, 8 May 2026 (UTC)

After looking at a sample of listed pages, many of the "no discernible vote" items may be Bobby nominating pages for deletion - if you're reading the output of afdstats, this is not true. Noms are in the table as Delete (Nom). "No discernible vote" items are ones where they edited the page but did not vote. For example, is Bobby adding a deletion sorting category. ~ A412 ^talk! 22:25, 8 May 2026 (UTC)

Ah, thanks for the clarification! It's unclear from the stats what is meant by "no discernible votes", and based on the number of noms, I wondered if this was the case. Significa liberdade (she/her) (talk) 00:42, 9 May 2026 (UTC)

Speedy deletion tagging: Bobby Cohn's CSD log dates back to October 2023, so I will analyse the full log. Each month has a variable number of nominations, ranging from 0 to around 60; there are around 770 entries in total. The overwhelming majority of listings are red links; those that are not are mostly either self-declines or recreations following deletion, although there are a few declines, mostly concentrated around the earlier years (, , , , , , , , , ). By my count, these constitute around 1.5% of the log; this percentage decreases to around 1% when you consider only the last year or so. Both of these figures are well within what I'd consider an "acceptable range". I'd also like to highlight a few entries that stood out: and show a willingness to reconsider in light of improvement, and shows self-critique and the ability to re-evaluate a decision. Also, after tagging the draftified redirect from Philip Egerton (died 1698) for R2, Bobby Cohn investigated further and had the integrity to move the draft back himself after realising it was too old (). All of these are qualities that I value in administrators; overall, this log is a good reflection on the candidate. UpTheOctave! • 8^va? 13:12, 9 May 2026 (UTC)