Re: Wikipedia at 25: A Wake-Up Call (essay)
WereSpielChequers via Wikimedia-l <[email protected]> Tue, 7 Jul 2026 08:46:07 +0100
| Newsgroups | gmane.org.wikimedia.foundation |
|---|---|
| Message-ID | <CAAanWP2NUtdqxGUeDXKixT=SqzQYG_i4SDfUumpdP3jUeZsArQ@mail.gmail.com> |
--===============5045622541147132452== Content-Type: multipart/alternative; boundary="000000000000371b7106560096e3" --000000000000371b7106560096e3 Content-Type: text/plain; charset="UTF-8" Content-Transfer-Encoding: quoted-printable James, I'm sorry to hear what happened with the medical articles. My own experience of resolving easily confused words, and of asking what various terms and phrases mean is that if you either make a minor tweak or ask what something means the first is rarely reversed and the second often results in a fix. But making a small change to a lot of articles is I'm sure easier than making a big change to a small number of articles. Perhaps a way to scale up would be to get consensus to deprecate some longwinded expressions where there are clearer alternatives. WSC On Sun, 5 Jul 2026 at 15:23, James Heilman <[email protected]> wrote: > Yes, writing the leads in easier to understand language would be > excellent. But this is definitely an uphill battle and anyone who tries m= ay > be confronted with alot of hate. Writing in easier to understand language > is called "dumbing down Wikipedia" by some and certain folks will go out = of > their way to prevent simplified leads from occurring. We did succeed for > healthcare content for a time per this paper: > > > https://en.wikiversity.org/wiki/WikiJournal_of_Medicine/Readability_of_En= glish_Wikipedia%27s_health_information_over_time > > But life is simply too short, and thus we forked to MDWiki.org for > simplified leads. > James > > On Sun, Jul 5, 2026 at 6:32=E2=80=AFAM WereSpielChequers via Wikimedia-l = < > [email protected]> wrote: > >> I'm not sure how useful my perspective is on this one, I'm a monolingual >> Brit who doesn't use AI myself, and while I have used a smartphone to ta= ke >> pictures on Commons I suspect over 99.99%, maybe 99.999% of my global ed= its >> have been PC or tablet based rather than from a smartphone. >> >> As Christophe has pointed out, AI is taking our audience. I'd add that w= e >> have had community growth issues for more than a decade, partly because = for >> smartphone users we are almost a broadcast medium. Optimising the mobile >> platform, and to some extent new skins on the PC platform, for readers >> rather than editors seems to have been WMF strategy for more than a deca= de. >> It succeeded in keeping our audience for a while, but at the price of >> restricting community growth and making us ever more dependent on aging = PC >> users in the global north. >> >> In a more growth oriented model we would recruit readers to become >> editors by incrementally changing their user experience, adding more use= ful >> editing features as they edit more. >> >> What I'm not sure is how much the deprioritising of Wikipedia content in >> favour of AI is a side effect of the big tech companies investing in AI,= or >> to what extent there are operators out there who see us as either a >> competitor or as a challenge to their alternative facts. I suspect it is= a >> bit of all three. >> >> Part of the solution is going to revolve around copyright. Both in terms >> of the WMF, or some other collective representation of Wikimedians, >> reminding AI operators of the SA and BY implications of CC-BY-SA. But al= so >> for the future, our potential role as a bridge between paywalled sites a= nd >> the free parts of the internet. There are still publications out there t= hat >> employ journalists and don't give away their content for free. Those of = us >> who have subscriptions to such publications give Wikipedia an advantage >> over AI content, that advantage isn't going to go away soon, and could >> increase if the WMF and Chapters invest more in the Wikipedia library an= d >> similar routes to make good offline sources available to Wikimedians. As= a >> side effect of this, I think we should do more outreach to the publishin= g >> industry as to what our criteria are for something to be considered a >> reliable source, and what changes certain sites would need to do for som= e >> or all of their content to meet our standards. >> >> One of the advantages that AI has or can have over us is in talking a >> language that people understand. I remember in the 2009 Strategy discuss= ion >> talking about following the Chinese Wikipedia system and making American >> English, Canadian English, Indian English and English as different views= of >> one Wikipedia. I don't know if that would also work for Portugal and Bra= zil >> and other languages that have become semi detached. But it is one change= we >> could fairly easily make. A more difficult one is reading ages. Apologie= s >> to other languages if this is just an English thing, but I get the >> impression that EN Wiki is written for a reading age that only a small >> minority of even native English speakers have achieved, whilst AI at lea= st >> tries to be understood by anyone with a day to day fluency in English. >> Lowering the reading age of our content, especially of the lede, would b= e a >> lot of work, but it is the sort of work that we could explain at U3A eve= nts >> or retirement fairs. We've long needed an alternative entry level tasks = as >> typos and vandalism get ever rarer on Wiki. >> >> There is a bigger cultural change to the movement that we need to start >> preparing for. Since the earliest days of the project, long before I got >> involved, our first level divide has been by language. As AI translation >> improves, we need to be aware that some barriers are coming down, we may >> find ourselves organising more by subject area, or find that some of our >> sources are reliable in their host language and less so when mirrored in= to >> other languages. Before Wikidata I was involved in some cross wiki worki= ng >> with the death anomaly project. I predict that this is going to be a gro= wth >> area for us as we start getting criticism for different language versio= ns >> of Wikipedia covering the same event from radically different perspectiv= es. >> >> We are living in interesting times. >> >> WereSpielChequers (=CB=88d=CA=92=C9=92n=C9=99=CE=B8=C9=99n >> <https://en.wikipedia.org/wiki/Help:IPA/English>) >> >> >> >> >> _______________________________________________ >> 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/Y55SJYOGQQIODKKA7QPLPEVRGEB6H6JC/ >> To unsubscribe send an email to wikimedia-l-leave-RusutVdil2icGmH+5r0DM0B+6BGkLq7r@public.gmane.org > > > > -- > James Heilman > MD, CCFP-EM, Wikipedian > --000000000000371b7106560096e3 Content-Type: text/html; charset="UTF-8" Content-Transfer-Encoding: quoted-printable <div dir=3D"ltr"><div>James,=C2=A0 I'm sorry to hear what happened with= the medical articles. My own experience of resolving easily confused words= , and of asking what various terms and phrases mean is that if you either m= ake a minor tweak or ask what something means the first is rarely reversed = and the second often results in a fix. But making a small change to a lot o= f articles is I'm sure easier than making a big change to a small numbe= r of articles.</div><div><br></div><div>Perhaps a way to scale up would be = to get consensus to deprecate some longwinded=C2=A0expressions where there = are clearer alternatives.</div><div><br></div><div>WSC</div><br><div class= =3D"gmail_quote gmail_quote_container"><div dir=3D"ltr" class=3D"gmail_attr= ">On Sun, 5 Jul 2026 at 15:23, James Heilman <<a href=3D"mailto:jmh649@g= mail.com">[email protected]</a>> wrote:<br></div><blockquote class=3D"gma= il_quote" style=3D"margin:0px 0px 0px 0.8ex;border-left:1px solid rgb(204,2= 04,204);padding-left:1ex"><div dir=3D"ltr">Yes, writing the leads in easier= to understand language would be excellent. But this is definitely=C2=A0an = uphill battle and anyone who tries may be confronted with alot of hate. Wri= ting=C2=A0in easier to understand=C2=A0language is called "dumbing dow= n Wikipedia" by some and certain folks will go out of their way to pre= vent simplified leads from occurring. We did succeed for healthcare content= for a time per this paper:<div><br></div><div><a href=3D"https://en.wikive= rsity.org/wiki/WikiJournal_of_Medicine/Readability_of_English_Wikipedia%27s= _health_information_over_time" target=3D"_blank">https://en.wikiversity.org= /wiki/WikiJournal_of_Medicine/Readability_of_English_Wikipedia%27s_health_i= nformation_over_time</a></div><div><br></div><div>But life is simply too sh= ort, and thus we forked to MDWiki.org for simplified leads.</div><div>James= </div></div><br><div class=3D"gmail_quote"><div dir=3D"ltr" class=3D"gmail_= attr">On Sun, Jul 5, 2026 at 6:32=E2=80=AFAM WereSpielChequers via Wikimedi= a-l <<a href=3D"mailto:[email protected]" target=3D"_blank= ">[email protected]</a>> wrote:<br></div><blockquote class= =3D"gmail_quote" style=3D"margin:0px 0px 0px 0.8ex;border-left:1px solid rg= b(204,204,204);padding-left:1ex"><div dir=3D"ltr">I'm not sure how usef= ul my perspective is on this one, I'm a monolingual Brit who doesn'= t use AI myself, and while I have used a smartphone to take pictures on Com= mons I suspect over 99.99%, maybe 99.999% of my global edits have been PC o= r tablet based=C2=A0rather than from a smartphone.<div><br></div><div>As Ch= ristophe has pointed out, AI is taking our audience. I'd add that we ha= ve had community growth issues for more than a decade, partly because for s= martphone users we are almost a broadcast medium. Optimising the mobile pla= tform, and to some extent new skins on the PC platform, for readers rather = than editors seems to have been WMF strategy for more than a decade. It suc= ceeded in keeping our audience for a while, but at the price of restricting= community growth and making us ever more dependent on aging PC users in th= e global north.</div><div><br></div><div>In a more growth oriented model we= would recruit readers to become editors by incrementally changing their us= er experience, adding more useful editing features=C2=A0as they edit more.<= /div><div><br></div><div>What I'm not sure is how much the deprioritisi= ng=C2=A0of Wikipedia=C2=A0content in favour of AI is a side effect of the b= ig tech companies investing in AI, or to what extent there are operators ou= t there who see us as either a competitor or as a challenge to their altern= ative facts. I suspect it is a bit of all three.=C2=A0</div><div><br></div>= <div>Part of the solution is going to revolve around copyright. Both in ter= ms of the WMF, or some other collective representation of Wikimedians, remi= nding AI operators of the SA and BY implications of CC-BY-SA. But also for = the future, our potential role as a bridge between paywalled sites and the = free parts of the internet. There are still publications out there that emp= loy journalists and don't give away their content for free. Those of us= who have subscriptions to such publications give Wikipedia an advantage ov= er AI content, that advantage isn't going to go away soon, and could in= crease if the WMF and Chapters invest more in the Wikipedia library and sim= ilar routes to make good offline sources available to Wikimedians. As a sid= e effect of this, I think we should do more outreach to the publishing indu= stry as to what our criteria are for something to be considered a reliable = source, and what changes certain sites would need to do for some or all of = their content to meet our standards.</div><div><br></div><div>One of the ad= vantages that AI has or can have over us is in talking a language that peop= le understand. I remember in the 2009 Strategy discussion talking about fol= lowing the Chinese Wikipedia system and making American English, Canadian E= nglish, Indian English and English as different views of one Wikipedia. I d= on't know if that would also work for Portugal and Brazil and other lan= guages that have become semi detached. But it is one change we could fairly= easily make. A more difficult one is reading ages. Apologies to other lang= uages if this is just an English thing, but I get the impression that EN Wi= ki is written for a reading age that only a small minority of even native E= nglish speakers have achieved, whilst AI at least tries to be understood by= anyone with a day to day fluency in English. Lowering the reading age of o= ur content, especially of the lede, would be a lot of work, but it is the s= ort of work that we could explain=C2=A0at U3A events or retirement fairs. W= e've long needed an alternative entry level tasks as typos and vandalis= m get ever rarer on Wiki.</div><div><br></div><div>There is a bigger cultur= al change to the movement that we need to start preparing for. Since the ea= rliest days of the project, long before I got involved, our first level div= ide has been by language. As AI translation improves, we need to be aware t= hat some barriers are coming down, we may find ourselves organising more by= subject area, or find that some of our sources are reliable in their host = language and less so when mirrored into other languages. Before Wikidata I = was involved in some cross wiki working with the death anomaly project. I p= redict that this is going to be a growth area for us as we start getting=C2= =A0 criticism for different language versions of Wikipedia covering the sam= e event from radically different perspectives.=C2=A0</div><div><br></div><d= iv>We are living in interesting times.</div><div><br></div><div>WereSpielCh= equers (<a rel=3D"mw:WikiLink" href=3D"https://en.wikipedia.org/wiki/Help:I= PA/English" title=3D"Help:IPA/English" style=3D"color:rgb(0,43,184);backgro= und:none rgb(248,249,250);font-family:sans-serif;font-size:11.176px;text-de= coration-line:none" target=3D"_blank"><span style=3D"border-bottom:1px dott= ed"><span title=3D"/=CB=88/: primary stress follows">=CB=88</span><span tit= le=3D"/d=CA=92/: 'j' in 'jam'">d=CA=92</span><span title=3D= "/=C9=92/: 'o' in 'body'">=C9=92</span><span title=3D"'= n' in 'nigh'">n</span><span title=3D"/=C9=99/: 'a' in &= #39;about'">=C9=99</span><span title=3D"/=CE=B8/: 'th' in '= thigh'">=CE=B8</span><span title=3D"/=C9=99n/: 'on' in 'but= ton'">=C9=99n</span></span></a>)</div><div><br></div><div>=C2=A0</div><= div><br></div><div><br></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/Y55SJYOGQQIODKKA7QPLPEVRGEB6H6JC/" r= el=3D"noreferrer" target=3D"_blank">https://lists.wikimedia.org/hyperkitty/= list/[email protected]/message/Y55SJYOGQQIODKKA7QPLPEVRGEB6H6= JC/</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> </blockquote></div></div> --000000000000371b7106560096e3-- --===============5045622541147132452== 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/KTOVVWAI642CHDC6JWZHF47SAMIGJYET/ To unsubscribe send an email to wikimedia-l-leave-RusutVdil2icGmH+5r0DM0B+6BGkLq7r@public.gmane.org --===============5045622541147132452==--