Re: Do we even know what a good AI Optimization strategy would be for Wikipedia? Re: Re: Google Zero is coming [was Re: Wikipedia at 25: A Wake-Up Call (essay)]
James Heilman via Wikimedia-l <[email protected]> Sat, 11 Jul 2026 16:31:17 +0200
| Newsgroups | gmane.org.wikimedia.foundation |
|---|---|
| Message-ID | <CAF1en7W7n_tbPBPnkATHAbWqeLOC2x-9Lc1HLSBC7D2djF=wfg@mail.gmail.com> |
--===============4337771092161704867== Content-Type: multipart/alternative; boundary="0000000000007ebdf4065656b6f1" --0000000000007ebdf4065656b6f1 Content-Type: text/plain; charset="UTF-8" Content-Transfer-Encoding: quoted-printable English Wikipedia is a conservative organization / project as are many other large versions of Wikipedia. Many stakeholders need to be convinced / brought on board to make even relatively minor changes. Innovating within smaller versions of Wikipedia or in other projects outside Wikipedia is much easier. And successes can occasionally be brought into the larger Wikipedias such as we did with Our World in Data interactive graphs... https://en.wikipedia.org/wiki/Wheat#Production_and_consumption We have now added 100s of these in various languages. And they are getting thousands of plays a day. James On Sat, Jul 11, 2026 at 3:28=E2=80=AFPM Charles Roberson via Wikimedia-l < [email protected]> wrote: > Wikimedia-I used to be a premium listserve of high minded ideas about how > to move the Foundation into the future. The past few months it has drifte= d > into a lot of whining about AI with no real effort to accomplish anything > or even plan to do so. > > - Charles > > On Sat, Jul 11, 2026 at 1:44=E2=80=AFAM Todd Allen via Wikimedia-l < > [email protected]> wrote: > >> We sure do know what a good AI strategy looks like for Wikipedia. No AI >> on Wikipedia. >> >> We have not succeeded by being FaceGramTwitTube. We have succeeded by no= t >> being like them. >> >> So, same here. No "latest and greatest". No AI on Wikipedia. Ever, for >> any reason, period. Wikipedia is written by people for people. >> >> Todd >> >> On Fri, Jul 10, 2026 at 11:09=E2=80=AFPM James Heilman via Wikimedia-l < >> [email protected]> wrote: >> >>> What has motivated me to spend time writing Wikipedia over the years is >>> writing for humans. The fact that the content is openly licensed and th= e >>> work is supported by an NGO is also key. >>> >>> Personally i do not feel any motivation to write primarily for trillion >>> dollar machines surrounded by venture capital folks hoping to make a >>> killing. If the machines want to adapt to human facing content sure. >>> >>> The approaches you mention is how Healthline succeeded, they basically >>> have dozens of articles covering the same topic just addressing it from= a >>> slightly different question. >>> >>> J >>> >>> >>> Sent from Gmail Mobile >>> >>> On Fri, Jul 10, 2026 at 21:24 Alex Stinson via Wikimedia-l < >>> [email protected]> wrote: >>> >>>> Forking this conversation, because I don't think we have a shared >>>> framing of what we are competing for in a "Google" Zero landscape domi= nated >>>> by ChatBot/AI search style RAG citations (i.e. >>>> https://en.wikipedia.org/wiki/Retrieval-augmented_generation). >>>> >>>> For the last 9 months, I've been examining how civil society content >>>> should be showing up in AI search, and there is a missing perspective = in >>>> our "we will build it and they will come" approach to Wikipedia. I don= 't >>>> think we can wait for the big tech companies or European regulatory bo= dies >>>> to adopt a different idea of how RAG should work. Here is my take fr= om >>>> what I have been engaging with in the AIO/AEO/SEO space: >>>> >>>> *Wikipedia doesn't have content that is SEO/AEO optimized* >>>> >>>> Part of the problem, even if we did have an MCP server is that the >>>> models (at least in my tracking), are pushing many citations away from >>>> "factual" websites, towards "authoritative" websites. This authoritati= ve >>>> content includes: >>>> >>>> - Expert original, analysis that makes strong claims based on facts >>>> (i.e. blog posts by authoritative companies or recently published >>>> ScienceDirect articles) >>>> - Content that has been updated recently, with the biggest "hot >>>> takes" (i.e. I have monitored a couple of prompt pools where citati= ons >>>> shift to newer content after 2-3 months) >>>> - Content that helps users make a decision between different >>>> choices (i.e. review websites, etc) >>>> >>>> >>>> This is following Google's longer-term push towards "human centered an= d >>>> useful" content (sometimes called E-E-A-T an abbreviation of experie= nce, >>>> expertise, authoritativeness, and trustworthiness, in SEO world). >>>> https://developers.google.com/search/docs/fundamentals/creating-helpfu= l-content >>>> >>>> >>>> To win in an AI optimization battle -- its less about Wikipedia doing >>>> well in the keyword search indexes that led to our content being visib= le >>>> (which is why we have a reputation as a "fact checking" website) and m= ore >>>> about "winning" in the criteria for what makes a good RAG citation ---= and >>>> our content format, is the exact opposite of the EAAT criteria: >>>> >>>> - Wikipedia is not authoriative, but rather points to other >>>> authorities >>>> - We ground our content in anonymity instead of named experts or >>>> instutional process/opinoin >>>> - We rarely do original analysis instead summarizing the experience >>>> and expertise of others, >>>> - Alot of our content is out of date, and self-aware of its gaps >>>> (i.e. maintenance tags), so also is likely to be undermining its ow= n >>>> trustworthiness >>>> >>>> >>>> *All the data points to us being used, but without an official roundup >>>> we are all talking in the dark about different assumed reputation loss= es* >>>> >>>> RAG unlike Google Search Indexing, seems to be using Wikipedia for a >>>> fraction of a fraction of responses, favoring these other kinds of >>>> sources: >>>> >>>> - Only 5% of AI overviews have Wikipedia in them: >>>> https://ahrefs.com/blog/most-cited-domains-ai-overviews/ >>>> - I have access to SERanking's corpus of prompt monitoring across 5 >>>> models (ChatGPT, Perplexity, Google Models and they suggest that in= May >>>> ~16% of prompts included Wikipdia, and in their most recent month (= June), >>>> ~13% of prompts. SERankings corpus is probably the # 4 or 5 in comm= ercial >>>> AIO data -- so could have gaps. >>>> - Studies from earlier in the year put Wikipedia at about ~13% of >>>> CHATGPT citations ( >>>> https://www.prnewswire.com/news-releases/wikipedia-and-reddit-now-d= rive-over-25-of-chatgpt-citations-in-the-us-new-5w-research-finds--wsj-nyt-= and-bloomberg-do-not-appear-in-the-top-20-302768339.html >>>> but chatgpt on average includes >20 sources in a response, compared= to >>>> googles 5-10 and doesn't expose it in the interface very well) >>>> - Comparable "top" Websites, like Youtube, Reddit, and LinkedIn >>>> tend to represent a greater % of content (in the SERanking data poo= l nearly >>>> 30% of responses had a Youtube Video cited for instance) >>>> - Domain specific citation pools have pretty significant >>>> differences in "which" sources are being called, with Wikipedia doi= ng well >>>> on some prompt pools: https://generativepulse.ai/report/ >>>> >>>> >>>> >>>> >>>> *RAG/AI search optimization focuses more on intent than keywords, and >>>> we aren't very effective at serving intent, and we don't know where ou= r >>>> optimization options are* >>>> What we need is an understanding of "which actual user reader behavior >>>> are we seeking to serve?". In the past we were extremely lazy, because >>>> keyword search always delivered Wikipedia as "a first". Now we need ou= r >>>> content to be more optimized for the kind of user curiosity driving th= eir >>>> use of a chatbot/search tool: >>>> >>>> - What percentage of prompts or AI searches are informational vs >>>> opinion forming? Are we even a competitor for grounding opinion bas= ed >>>> questions or only the informational ones? >>>> - How many of the interactions are two or three steps down a chain >>>> of more "specific" interactions with the chatbot and thus no longer= need >>>> "general knowledge" information from Wikipedia, but rather the kind= s of >>>> stuff that we rely on our citations to provide ? >>>> - How much are the AI companies optimizing for "sales" or >>>> "addiction" rather than for leading users to reliable content? (I w= as >>>> tracking a series of informational topics about food that (on ChatG= PT and >>>> Google), kept wanting me to continue the conversation by *inviting >>>> me to go to local hamburger restraunts)*. Do we even have a >>>> reasonable chance to be in those searches? >>>> - How much is geolocation forcing more and more responses into >>>> "local" sources rather than "global" websites? In one dataset I tra= cked, in >>>> Global South countries citations were overwhelmingly to Facebook an= d >>>> Instagram despite more authoritative academic, news and Wikipedia-t= ype >>>> sites in the same searches from the UK. >>>> >>>> >>>> *We may need to radically change the "readable signals" on our content >>>> pages, meaning changing the Manual of Style, Editing Practices, and AI >>>> enabled enrichment.* >>>> >>>> If we are trying to market Wikipedia's content into AI interfaces, we >>>> also can't do what most AI optimization/marketing agencies would sugge= st: >>>> writing listicle/FAQ type content that closely matches the user-querie= s >>>> that folks are giving the IA models (i.e. analysis like: >>>> https://neilpatel.com/marketing-stats/trust-signals-ai-engines-reward-= most/ >>>> ). >>>> >>>> We would then have to experiment with other content types, that _no >>>> longer look like the encyclopedia_. Or we would need to be reconfiguri= ng >>>> the Encyclopedic content to expose enrichments to paragraphs or sectio= ns >>>> within the encyclopedia that pretty radically change editorial assupm= tions >>>> and our Manual of Style (i.e. instead of simple 1-2 word section headi= ngs, >>>> like "History" we may need intent-focused headings like "What is the >>>> history of [x topic]?). >>>> >>>> If we want to compete in the shifting AI search landscape -- we would >>>> need a lot more data from the Foundation on where we are succeeding or= not, >>>> and then consider *_radically different_ *ways of exposing our content >>>> in terms of treating RAG systems as a user that needs correct paths to >>>> Wikipedia pages. >>>> >>>> However, this doesn't necessarily need to change the *human reader >>>> experience *, but would need to be about configuring the content >>>> (beyond an MCP server or Enterpise APIs) *for an AI audience/consumer >>>> experience -- *which I haven't seen addressed in any WMF publications >>>> or community conversations. Without a firm theory of "What kind of con= sumer >>>> is an AI search agent/RAG index?" and "How does our content need to se= rve >>>> that AI audience?" the editing community won't be able to adjust its >>>> editing practices or weigh in on feature recommendations that make our >>>> content "AI useful". >>>> >>>> As I have written elsewhere, I think there is a inherent audience for >>>> editing/using the Wikis organically: >>>> https://en.wikipedia.org/wiki/Wikipedia:Wikipedia_Signpost/2026-06-21/= Op-ed >>>> -- but its a different question than competing "with other information >>>> sources" for AI as an audience. >>>> >>>> >>>> >>>> >>>> >>>> On Fri, Jul 10, 2026 at 2:30=E2=80=AFPM Steven Walling via Wikimedia-l= < >>>> [email protected]> wrote: >>>> >>>>> >>>>> >>>>> On Fri, Jul 10, 2026 at 9:52=E2=80=AFAM Erik Moeller via Wikimedia-l = < >>>>> [email protected]> wrote: >>>>> >>>>>> On Fri, Jul 10, 2026 at 5:59=E2=80=AFPM James Heilman via Wikimedia-= l >>>>>> <[email protected]> wrote: >>>>>> >>>>>> > Yah a search engine that actually gives real references that >>>>>> supports the statements in question would be amazing. >>>>>> >>>>>> Almost like .. a Knowledge Engine. ;-) >>>>>> >>>>> >>>>> Is the WMF building an MCP server to connect Wikipedia and Wikidata >>>>> directly to Gemini, Claude, and ChatGPT? This is a more lightweight, >>>>> backdoor way to leverage the audience of those platforms but present >>>>> structured outputs to AI chats for users based on Wikimedia knowledge= . If >>>>> we did so, we could present citations within the returned responses t= o >>>>> users, and those platforms make it transparent to the user when they = are >>>>> calling a particular tool. >>>>> >>>>> Steven Walling >>>>> >>>>> Sadly, the only realistic path I see there would be through >>>>>> acquisition, and even if that was financially feasible, you'd begin = by >>>>>> inheriting a lot of corporate practices that aren't really consisten= t >>>>>> with Wikimedia values. >>>>>> >>>>>> But perhaps there is a middle ground where Wikimedia seeks to define >>>>>> more clearly the terms of engagement that it wants with search engin= es >>>>>> (clear attribution, clear and correct references, calls-to-edit, >>>>>> etc.), and then finds and recognizes search partners who implement >>>>>> those. To Luis' point, that need not be done by WMF. >>>>>> >>>>>> Warmly, >>>>>> >>>>>> Erik >>>>>> _______________________________________________ >>>>>> Wikimedia-l mailing list -- [email protected], >>>>>> guidelines at: >>>>>> https://meta.wikimedia.org/wiki/Mailing_lists/Guidelines and >>>>>> https://meta.wikimedia.org/wiki/Wikimedia-l >>>>>> Public archives at >>>>>> https://lists.wikimedia.org/hyperkitty/list/[email protected]= dia.org/message/HNDPDZBBDMILF7WGMUBVJVIAZYQ7OXOS/ >>>>>> To unsubscribe send an email to wikimedia-l-leave-RusutVdil2icGmH+5r0DM8Xa4x6EXUF0@public.gmane.org= g >>>>> >>>>> _______________________________________________ >>>>> Wikimedia-l mailing list -- [email protected], >>>>> guidelines at: >>>>> https://meta.wikimedia.org/wiki/Mailing_lists/Guidelines and >>>>> https://meta.wikimedia.org/wiki/Wikimedia-l >>>>> Public archives at >>>>> https://lists.wikimedia.org/hyperkitty/list/[email protected]= ia.org/message/IUDAKJRN5EQT5CCWEEYQXKDSVXCDBXR6/ >>>>> To unsubscribe send an email to wikimedia-l-leave-RusutVdil2icGmH+5r0DM0B+6BGkLq7r@public.gmane.org >>>> >>>> _______________________________________________ >>>> Wikimedia-l mailing list -- [email protected], >>>> guidelines at: https://meta.wikimedia.org/wiki/Mailing_lists/Guideline= s >>>> and https://meta.wikimedia.org/wiki/Wikimedia-l >>>> Public archives at >>>> https://lists.wikimedia.org/hyperkitty/list/[email protected]= a.org/message/YODBKTCLA2XJC24Y23I2SUN3PUYA4QXE/ >>>> To unsubscribe send an email to wikimedia-l-leave-RusutVdil2icGmH+5r0DM0B+6BGkLq7r@public.gmane.org >>> >>> _______________________________________________ >>> Wikimedia-l mailing list -- [email protected], guidelines >>> at: https://meta.wikimedia.org/wiki/Mailing_lists/Guidelines and >>> https://meta.wikimedia.org/wiki/Wikimedia-l >>> Public archives at >>> https://lists.wikimedia.org/hyperkitty/list/[email protected]= .org/message/5MPMZGZ2MXLLHER3NTNUS6KHIMAXIIBD/ >>> To unsubscribe send an email to wikimedia-l-leave-RusutVdil2icGmH+5r0DM0B+6BGkLq7r@public.gmane.org >> >> _______________________________________________ >> Wikimedia-l mailing list -- [email protected], guidelines >> at: https://meta.wikimedia.org/wiki/Mailing_lists/Guidelines and >> https://meta.wikimedia.org/wiki/Wikimedia-l >> Public archives at >> https://lists.wikimedia.org/hyperkitty/list/[email protected]= org/message/QAGHRXXTGSXKVFOALBRQHP5G27NUIFDD/ >> To unsubscribe send an email to wikimedia-l-leave-RusutVdil2icGmH+5r0DM0B+6BGkLq7r@public.gmane.org > > _______________________________________________ > Wikimedia-l mailing list -- [email protected], guidelines > at: https://meta.wikimedia.org/wiki/Mailing_lists/Guidelines and > https://meta.wikimedia.org/wiki/Wikimedia-l > Public archives at > https://lists.wikimedia.org/hyperkitty/list/[email protected]= rg/message/6RKSJMWMVUHUUVV7WGKPOCOHCN4NBQK4/ > To unsubscribe send an email to wikimedia-l-leave-RusutVdil2icGmH+5r0DM0B+6BGkLq7r@public.gmane.org --=20 James Heilman MD, CCFP-EM, Wikipedian --0000000000007ebdf4065656b6f1 Content-Type: text/html; charset="UTF-8" Content-Transfer-Encoding: quoted-printable <div dir=3D"ltr">English Wikipedia is a conservative organization / project= as are many other large versions of Wikipedia. Many stakeholders need to b= e convinced / brought on board to make even relatively minor changes. Innov= ating within smaller versions of Wikipedia or in other projects outside Wik= ipedia is much easier. And successes can occasionally=C2=A0be brought into = the larger Wikipedias such as we did with Our World in Data interactive gra= phs...<div><br></div><div><a href=3D"https://en.wikipedia.org/wiki/Wheat#Pr= oduction_and_consumption">https://en.wikipedia.org/wiki/Wheat#Production_an= d_consumption</a></div><div><br></div><div>We have now added 100s of these = in various languages. And they are getting thousands of plays a day.</div><= div><br></div><div>James</div></div><br><div class=3D"gmail_quote gmail_quo= te_container"><div dir=3D"ltr" class=3D"gmail_attr">On Sat, Jul 11, 2026 at= 3:28=E2=80=AFPM Charles Roberson via Wikimedia-l <<a href=3D"mailto:wik= [email protected]">[email protected]</a>> wrote= :<br></div><blockquote class=3D"gmail_quote" style=3D"margin:0px 0px 0px 0.= 8ex;border-left:1px solid rgb(204,204,204);padding-left:1ex"><div dir=3D"lt= r">Wikimedia-I used to be a premium listserve=C2=A0of high minded ideas abo= ut how to move the Foundation into the future. The past few months it has d= rifted into a lot of whining about AI with no real effort to accomplish any= thing or even plan to do so.<div><br></div><div>=C2=A0- Charles</div></div>= <br><div class=3D"gmail_quote"><div dir=3D"ltr" class=3D"gmail_attr">On Sat= , Jul 11, 2026 at 1:44=E2=80=AFAM Todd Allen via Wikimedia-l <<a href=3D= "mailto:[email protected]" target=3D"_blank">wikimedia-l@list= s.wikimedia.org</a>> wrote:<br></div><blockquote class=3D"gmail_quote" s= tyle=3D"margin:0px 0px 0px 0.8ex;border-left:1px solid rgb(204,204,204);pad= ding-left:1ex"><div dir=3D"ltr"><div>We sure do know what a good AI strateg= y looks like for Wikipedia. No AI on Wikipedia.</div><div><br></div><div>We= have not succeeded by being FaceGramTwitTube. We have succeeded by not bei= ng like them.</div><div><br></div><div>So, same here. No "latest and g= reatest". No AI on Wikipedia. Ever, for any reason, period. Wikipedia = is written by people for people.</div><div><br></div><div>Todd</div></div><= br><div class=3D"gmail_quote"><div dir=3D"ltr" class=3D"gmail_attr">On Fri,= Jul 10, 2026 at 11:09=E2=80=AFPM James Heilman via Wikimedia-l <<a href= =3D"mailto:[email protected]" target=3D"_blank">wikimedia-l@l= ists.wikimedia.org</a>> wrote:<br></div><blockquote class=3D"gmail_quote= " style=3D"margin:0px 0px 0px 0.8ex;border-left:1px solid rgb(204,204,204);= padding-left:1ex"><div dir=3D"auto">What has motivated me to spend time wri= ting Wikipedia over the years is writing for humans. The fact that the cont= ent is openly licensed and the work is supported by an NGO is also key.</di= v><div dir=3D"auto"><br></div><div dir=3D"auto">Personally i do not feel an= y motivation to write primarily for trillion dollar machines surrounded by = venture capital folks hoping to make a killing. If the machines want to ada= pt to human facing content sure.</div><div dir=3D"auto"><br></div><div dir= =3D"auto">The approaches you mention is how Healthline succeeded, they basi= cally have dozens of articles covering the same topic just addressing it fr= om a slightly different question.=C2=A0</div><div dir=3D"auto"><br></div><d= iv dir=3D"auto">J</div><div><br clear=3D"all"><br clear=3D"all"><div><div d= ir=3D"ltr" class=3D"gmail_signature">Sent from Gmail Mobile</div></div></di= v><div><br><div class=3D"gmail_quote"><div dir=3D"ltr" class=3D"gmail_attr"= >On Fri, Jul 10, 2026 at 21:24 Alex Stinson via Wikimedia-l <<a href=3D"= mailto:[email protected]" target=3D"_blank">wikimedia-l@lists= .wikimedia.org</a>> wrote:<br></div><blockquote class=3D"gmail_quote" st= yle=3D"margin:0px 0px 0px 0.8ex;border-left:1px solid rgb(204,204,204);padd= ing-left:1ex"><div dir=3D"ltr"><div dir=3D"ltr">Forking this conversation, = because I don't think we have a shared framing of what we are competing= for in a "Google" Zero landscape dominated by ChatBot/AI search = style RAG citations (i.e. <a href=3D"https://en.wikipedia.org/wiki/Retrieva= l-augmented_generation" target=3D"_blank">https://en.wikipedia.org/wiki/Ret= rieval-augmented_generation</a>).<br><br>For the last 9 months, I've be= en examining how civil society content should be showing up in AI search, a= nd there is a missing perspective in our "we will build it and they wi= ll come" approach to Wikipedia. I don't think we can wait for the = big tech companies or European regulatory bodies to adopt a different idea = of how RAG should work.=C2=A0 =C2=A0Here is my take from what I have been e= ngaging with in the AIO/AEO/SEO space:<br><br><b>Wikipedia doesn't have= content that is SEO/AEO optimized</b><br><br>Part of the problem, even if = we did have an MCP server is that the models=C2=A0(at least in my tracking)= , are pushing many citations away from "factual" websites, toward= s "authoritative" websites. This authoritative content includes:= =C2=A0<br><ul><li>Expert original, analysis that makes strong claims based = on facts (i.e. blog posts by authoritative companies or recently published = ScienceDirect articles)</li><li>Content that has been updated recently, wit= h the biggest "hot takes" (i.e. I have monitored a couple of prom= pt pools where citations shift to newer content after 2-3 months)</li><li>C= ontent that helps users make a decision between different choices (i.e. rev= iew websites, etc)</li></ul><br>This is following Google's longer-term = push towards "human centered and useful" content (sometimes calle= d=C2=A0=C2=A0E-E-A-T an abbreviation of=C2=A0 experience, expertise, author= itativeness, and trustworthiness, in SEO world).=C2=A0<a href=3D"https://de= velopers.google.com/search/docs/fundamentals/creating-helpful-content" targ= et=3D"_blank">https://developers.google.com/search/docs/fundamentals/creati= ng-helpful-content</a>=C2=A0<br><br>To win in an AI optimization battle -- = its less about Wikipedia doing well in the keyword search indexes that led = to our content being visible (which is why we have a reputation as a "= fact checking" website) and more about "winning" in the crit= eria for what makes a good RAG citation --- and our content format, is the = exact opposite of the EAAT criteria:=C2=A0<br><ul><li>Wikipedia is not auth= oriative, but rather points to other authorities</li><li>We ground our cont= ent in anonymity instead of named experts or instutional process/opinoin</l= i><li>We rarely do original analysis instead summarizing the experience and= expertise of others,=C2=A0</li><li>Alot of our content is out of date, and= self-aware of its gaps (i.e. maintenance tags), so also is likely to be un= dermining its own trustworthiness</li></ul><br><b>All the data points to us= being used, but without an official roundup we are all=C2=A0talking in the= dark about different assumed reputation losses</b><br><br>RAG unlike Googl= e Search Indexing, seems to be using Wikipedia for a fraction of a fraction= of responses, favoring these other kinds of sources:=C2=A0=C2=A0<br><ul><l= i>Only 5% of AI overviews have Wikipedia in them:=C2=A0<a href=3D"https://a= hrefs.com/blog/most-cited-domains-ai-overviews/" target=3D"_blank">https://= ahrefs.com/blog/most-cited-domains-ai-overviews/</a></li><li>I have access = to SERanking's corpus of prompt monitoring across 5 models (ChatGPT, Pe= rplexity, Google Models and they suggest that in May ~16% of prompts includ= ed Wikipdia, and in their most recent month (June), ~13% of prompts. SERank= ings corpus is probably the # 4 or 5 in commercial AIO data -- so could hav= e gaps.</li><li>Studies from earlier in the year put Wikipedia at about ~13= % of CHATGPT citations ( <a href=3D"https://www.prnewswire.com/news-release= s/wikipedia-and-reddit-now-drive-over-25-of-chatgpt-citations-in-the-us-new= -5w-research-finds--wsj-nyt-and-bloomberg-do-not-appear-in-the-top-20-30276= 8339.html" target=3D"_blank">https://www.prnewswire.com/news-releases/wikip= edia-and-reddit-now-drive-over-25-of-chatgpt-citations-in-the-us-new-5w-res= earch-finds--wsj-nyt-and-bloomberg-do-not-appear-in-the-top-20-302768339.ht= ml</a> but chatgpt on average includes >20 sources in a response, compar= ed to googles 5-10 and doesn't expose it in the interface very well)=C2= =A0</li><li>Comparable "top" Websites, like Youtube, Reddit, and = LinkedIn tend to represent a greater % of content (in the SERanking data po= ol nearly 30% of responses had a Youtube Video cited for instance)</li><li>= Domain specific citation pools have pretty significant differences in "= ;which" sources are being called, with Wikipedia doing well on some pr= ompt pools:=C2=A0<a href=3D"https://generativepulse.ai/report/" target=3D"_= blank">https://generativepulse.ai/report/</a></li></ul><br><br><b>RAG/AI se= arch optimization focuses more on intent than keywords, and we aren't v= ery effective at serving intent, and we don't know where our optimizati= on options are<br></b><br>What we need is an understanding of "which a= ctual user reader behavior are we seeking to serve?". In the past we w= ere extremely lazy, because keyword search always delivered Wikipedia as &q= uot;a first". Now we need our content to be more optimized for the kin= d of user curiosity driving their use of a chatbot/search tool:=C2=A0<br><u= l><li>What percentage of prompts or AI searches are informational vs opinio= n forming? Are we even a competitor for grounding opinion based questions o= r only the informational ones?</li><li>How many of the interactions are two= or three steps down a chain of more "specific" interactions with= the chatbot and thus no longer need "general knowledge" informat= ion from Wikipedia, but rather the kinds of stuff that we rely on our citat= ions to provide ?=C2=A0</li><li>How much are the AI companies optimizing fo= r "sales" or "addiction" rather than for leading users = to reliable content? (I was tracking a series of informational topics about= food that (on ChatGPT and Google), kept wanting me to continue the convers= ation by <i>inviting me to go to local hamburger restraunts)</i>. Do we eve= n have a reasonable chance to be in those searches?</li><li>How much is geo= location forcing more and more responses into "local" sources rat= her than "global" websites? In one dataset I tracked, in Global S= outh countries citations were overwhelmingly to Facebook and Instagram desp= ite more authoritative academic, news and Wikipedia-type sites in the same = searches from the UK.</li></ul><br><b>We may need to radically change the &= quot;readable signals" on our content pages, meaning changing the Manu= al of Style, Editing Practices, and AI enabled enrichment.</b><br><br>If we= are trying to market Wikipedia's content into AI interfaces, we also c= an't do what most AI optimization/marketing agencies would suggest: wri= ting listicle/FAQ type content that closely matches the user-queries that f= olks are giving the IA models (i.e. analysis like: <a href=3D"https://neilp= atel.com/marketing-stats/trust-signals-ai-engines-reward-most/" target=3D"_= blank">https://neilpatel.com/marketing-stats/trust-signals-ai-engines-rewar= d-most/</a> ). <br><br>We would then have to experiment with other content = types, that _no longer look like the encyclopedia_. Or we would need to be = reconfiguring the Encyclopedic content to expose enrichments to paragraphs = or sections within the encyclopedia=C2=A0 that pretty radically change edit= orial assupmtions and our Manual of Style (i.e. instead of simple 1-2 word = section headings, like "History" we may need intent-focused headi= ngs like "What is the history of [x topic]?).=C2=A0<br><br>If we want = to compete in the shifting AI search landscape -- we would need a lot more = data from the Foundation on where we are succeeding or not, and then consid= er <i>_radically different_ </i>ways of exposing our content in terms of tr= eating RAG systems as a user that needs correct paths to Wikipedia pages.<b= r>=C2=A0<br>However, this doesn't necessarily need to change the <i>hum= an=C2=A0reader experience </i>, but would need to be about configuring the = content (beyond an MCP server or Enterpise APIs)=C2=A0<i>for an AI audience= /consumer experience -- </i>which I haven't seen addressed in any WMF p= ublications or community conversations. Without a firm theory of "What= kind of consumer is an AI search agent/RAG index?" and "How does= our content need to serve that AI audience?" the editing community wo= n't be able to adjust its editing practices or weigh in on feature reco= mmendations that make our content "AI useful".<br><br>As I have w= ritten elsewhere, I think there is a inherent audience for editing/using th= e Wikis organically:=C2=A0<a href=3D"https://en.wikipedia.org/wiki/Wikipedi= a:Wikipedia_Signpost/2026-06-21/Op-ed" target=3D"_blank">https://en.wikiped= ia.org/wiki/Wikipedia:Wikipedia_Signpost/2026-06-21/Op-ed</a> -- but its a = different question than competing "with other information sources"= ; for AI as an audience.<br><br><br><br><br></div><br><div class=3D"gmail_q= uote"><div dir=3D"ltr" class=3D"gmail_attr">On Fri, Jul 10, 2026 at 2:30=E2= =80=AFPM Steven Walling via Wikimedia-l <<a href=3D"mailto:wikimedia-l@l= ists.wikimedia.org" target=3D"_blank">[email protected]</a>&g= t; wrote:<br></div><blockquote class=3D"gmail_quote" style=3D"margin:0px 0p= x 0px 0.8ex;border-left:1px solid rgb(204,204,204);padding-left:1ex"><div d= ir=3D"ltr"><div dir=3D"ltr"><br></div><br><div class=3D"gmail_quote"><div d= ir=3D"ltr" class=3D"gmail_attr">On Fri, Jul 10, 2026 at 9:52=E2=80=AFAM Eri= k Moeller via Wikimedia-l <<a href=3D"mailto:[email protected]= .org" target=3D"_blank">[email protected]</a>> wrote:<br><= /div><blockquote class=3D"gmail_quote" style=3D"margin:0px 0px 0px 0.8ex;bo= rder-left:1px solid rgb(204,204,204);padding-left:1ex">On Fri, Jul 10, 2026= at 5:59=E2=80=AFPM James Heilman via Wikimedia-l<br> <<a href=3D"mailto:[email protected]" target=3D"_blank">wi= [email protected]</a>> wrote:<br> <br> > Yah a search engine that actually gives real references that supports = the statements in question would be amazing.<br> <br> Almost like .. a Knowledge Engine. ;-)<br></blockquote><div><br></div>Is th= e WMF building an MCP server to connect Wikipedia and Wikidata directly to = Gemini, Claude, and ChatGPT? This is a more lightweight, backdoor way to le= verage the audience of those platforms but present structured outputs to AI= chats for users based on Wikimedia knowledge.=C2=A0<span style=3D"backgrou= nd-color:transparent">If we did so, we could present citations within the r= eturned responses to users, and those platforms make it transparent to the = user when they are calling a particular tool.=C2=A0</span><span style=3D"ba= ckground-color:transparent">=C2=A0</span></div><div class=3D"gmail_quote"><= span style=3D"background-color:transparent"><br></span></div><div class=3D"= gmail_quote"><span style=3D"background-color:transparent">Steven Walling=C2= =A0</span></div><div class=3D"gmail_quote"><br><blockquote class=3D"gmail_q= uote" style=3D"margin:0px 0px 0px 0.8ex;border-left:1px solid rgb(204,204,2= 04);padding-left:1ex">Sadly, the only realistic path I see there would be t= hrough<br> acquisition, and even if that was financially feasible, you'd begin by<= br> inheriting a lot of corporate practices that aren't really consistent<b= r> with Wikimedia values.<br> <br> But perhaps there is a middle ground where Wikimedia seeks to define<br> more clearly the terms of engagement that it wants with search engines<br> (clear attribution, clear and correct references, calls-to-edit,<br> etc.), and then finds and recognizes search partners who implement<br> those. To Luis' point, that need not be done by WMF.<br> <br> Warmly,<br> <br> Erik<br> _______________________________________________<br> Wikimedia-l mailing list -- <a href=3D"mailto:[email protected]= rg" target=3D"_blank">[email protected]</a>, guidelines at: <= a href=3D"https://meta.wikimedia.org/wiki/Mailing_lists/Guidelines" rel=3D"= noreferrer" target=3D"_blank">https://meta.wikimedia.org/wiki/Mailing_lists= /Guidelines</a> and <a href=3D"https://meta.wikimedia.org/wiki/Wikimedia-l"= rel=3D"noreferrer" target=3D"_blank">https://meta.wikimedia.org/wiki/Wikim= edia-l</a><br> Public archives at <a href=3D"https://lists.wikimedia.org/hyperkitty/list/w= [email protected]/message/HNDPDZBBDMILF7WGMUBVJVIAZYQ7OXOS/" r= el=3D"noreferrer" target=3D"_blank">https://lists.wikimedia.org/hyperkitty/= list/[email protected]/message/HNDPDZBBDMILF7WGMUBVJVIAZYQ7OX= OS/</a><br> To unsubscribe send an email to <a href=3D"mailto:[email protected]= ikimedia.org" target=3D"_blank">wikimedia-l-leave-RusutVdil2icGmH+5r0DM0B+6BGkLq7r@public.gmane.org</a></= blockquote></div></div> _______________________________________________<br> Wikimedia-l mailing list -- <a href=3D"mailto:[email protected]= rg" target=3D"_blank">[email protected]</a>, guidelines at: <= a href=3D"https://meta.wikimedia.org/wiki/Mailing_lists/Guidelines" rel=3D"= noreferrer" target=3D"_blank">https://meta.wikimedia.org/wiki/Mailing_lists= /Guidelines</a> and <a href=3D"https://meta.wikimedia.org/wiki/Wikimedia-l"= rel=3D"noreferrer" target=3D"_blank">https://meta.wikimedia.org/wiki/Wikim= edia-l</a><br> Public archives at <a href=3D"https://lists.wikimedia.org/hyperkitty/list/w= [email protected]/message/IUDAKJRN5EQT5CCWEEYQXKDSVXCDBXR6/" r= el=3D"noreferrer" target=3D"_blank">https://lists.wikimedia.org/hyperkitty/= list/[email protected]/message/IUDAKJRN5EQT5CCWEEYQXKDSVXCDBX= R6/</a><br> To unsubscribe send an email to <a href=3D"mailto:[email protected]= ikimedia.org" target=3D"_blank">wikimedia-l-leave-RusutVdil2icGmH+5r0DM0B+6BGkLq7r@public.gmane.org</a></= blockquote></div></div> _______________________________________________<br> Wikimedia-l mailing list -- <a href=3D"mailto:[email protected]= rg" target=3D"_blank">[email protected]</a>, guidelines at: <= a href=3D"https://meta.wikimedia.org/wiki/Mailing_lists/Guidelines" rel=3D"= noreferrer" target=3D"_blank">https://meta.wikimedia.org/wiki/Mailing_lists= /Guidelines</a> and <a href=3D"https://meta.wikimedia.org/wiki/Wikimedia-l"= rel=3D"noreferrer" target=3D"_blank">https://meta.wikimedia.org/wiki/Wikim= edia-l</a><br> Public archives at <a href=3D"https://lists.wikimedia.org/hyperkitty/list/w= [email protected]/message/YODBKTCLA2XJC24Y23I2SUN3PUYA4QXE/" r= el=3D"noreferrer" target=3D"_blank">https://lists.wikimedia.org/hyperkitty/= list/[email protected]/message/YODBKTCLA2XJC24Y23I2SUN3PUYA4Q= XE/</a><br> To unsubscribe send an email to <a href=3D"mailto:[email protected]= ikimedia.org" target=3D"_blank">wikimedia-l-leave-RusutVdil2icGmH+5r0DM0B+6BGkLq7r@public.gmane.org</a></= blockquote></div></div> _______________________________________________<br> Wikimedia-l mailing list -- <a href=3D"mailto:[email protected]= rg" target=3D"_blank">[email protected]</a>, guidelines at: <= a href=3D"https://meta.wikimedia.org/wiki/Mailing_lists/Guidelines" rel=3D"= noreferrer" target=3D"_blank">https://meta.wikimedia.org/wiki/Mailing_lists= /Guidelines</a> and <a href=3D"https://meta.wikimedia.org/wiki/Wikimedia-l"= rel=3D"noreferrer" target=3D"_blank">https://meta.wikimedia.org/wiki/Wikim= edia-l</a><br> Public archives at <a href=3D"https://lists.wikimedia.org/hyperkitty/list/w= [email protected]/message/5MPMZGZ2MXLLHER3NTNUS6KHIMAXIIBD/" r= el=3D"noreferrer" target=3D"_blank">https://lists.wikimedia.org/hyperkitty/= list/[email protected]/message/5MPMZGZ2MXLLHER3NTNUS6KHIMAXII= BD/</a><br> To unsubscribe send an email to <a href=3D"mailto:[email protected]= ikimedia.org" target=3D"_blank">wikimedia-l-leave-RusutVdil2icGmH+5r0DM0B+6BGkLq7r@public.gmane.org</a></= blockquote></div> _______________________________________________<br> Wikimedia-l mailing list -- <a href=3D"mailto:[email protected]= rg" target=3D"_blank">[email protected]</a>, guidelines at: <= a href=3D"https://meta.wikimedia.org/wiki/Mailing_lists/Guidelines" rel=3D"= noreferrer" target=3D"_blank">https://meta.wikimedia.org/wiki/Mailing_lists= /Guidelines</a> and <a href=3D"https://meta.wikimedia.org/wiki/Wikimedia-l"= rel=3D"noreferrer" target=3D"_blank">https://meta.wikimedia.org/wiki/Wikim= edia-l</a><br> Public archives at <a href=3D"https://lists.wikimedia.org/hyperkitty/list/w= [email protected]/message/QAGHRXXTGSXKVFOALBRQHP5G27NUIFDD/" r= el=3D"noreferrer" target=3D"_blank">https://lists.wikimedia.org/hyperkitty/= list/[email protected]/message/QAGHRXXTGSXKVFOALBRQHP5G27NUIF= DD/</a><br> To unsubscribe send an email to <a href=3D"mailto:[email protected]= ikimedia.org" target=3D"_blank">wikimedia-l-leave-RusutVdil2icGmH+5r0DM0B+6BGkLq7r@public.gmane.org</a></= blockquote></div> _______________________________________________<br> Wikimedia-l mailing list -- <a href=3D"mailto:[email protected]= rg" target=3D"_blank">[email protected]</a>, guidelines at: <= a href=3D"https://meta.wikimedia.org/wiki/Mailing_lists/Guidelines" rel=3D"= noreferrer" target=3D"_blank">https://meta.wikimedia.org/wiki/Mailing_lists= /Guidelines</a> and <a href=3D"https://meta.wikimedia.org/wiki/Wikimedia-l"= rel=3D"noreferrer" target=3D"_blank">https://meta.wikimedia.org/wiki/Wikim= edia-l</a><br> Public archives at <a href=3D"https://lists.wikimedia.org/hyperkitty/list/w= [email protected]/message/6RKSJMWMVUHUUVV7WGKPOCOHCN4NBQK4/" r= el=3D"noreferrer" target=3D"_blank">https://lists.wikimedia.org/hyperkitty/= list/[email protected]/message/6RKSJMWMVUHUUVV7WGKPOCOHCN4NBQ= K4/</a><br> To unsubscribe send an email to <a href=3D"mailto:[email protected]= ikimedia.org" target=3D"_blank">wikimedia-l-leave-RusutVdil2icGmH+5r0DM0B+6BGkLq7r@public.gmane.org</a></= blockquote></div><div><br clear=3D"all"></div><div><br></div><span class=3D= "gmail_signature_prefix">-- </span><br><div dir=3D"ltr" class=3D"gmail_sign= ature"><div dir=3D"ltr"><div><div dir=3D"ltr"><div dir=3D"ltr"><div dir=3D"= ltr">James Heilman<br>MD, CCFP-EM, Wikipedian</div></div></div></div></div>= </div> --0000000000007ebdf4065656b6f1-- --===============4337771092161704867== Content-Type: text/plain; charset="us-ascii" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit Content-Disposition: inline _______________________________________________ Wikimedia-l mailing list -- [email protected], guidelines at: https://meta.wikimedia.org/wiki/Mailing_lists/Guidelines and https://meta.wikimedia.org/wiki/Wikimedia-l Public archives at https://lists.wikimedia.org/hyperkitty/list/[email protected]/message/FRGCURSOK3KPMT3UWZFOR6Z4TMZFE7KF/ To unsubscribe send an email to wikimedia-l-leave-RusutVdil2icGmH+5r0DM0B+6BGkLq7r@public.gmane.org --===============4337771092161704867==--