Speaker 00:00:02
Heather Ford: Josh Lim is what the media calls a Wikipedia warrior.
Speaker 00:00:07
Josh Lim: You don't live in a democracy without free access to information.
Speaker 00:00:13
Heather Ford: An editor trying to keep information accurate and open in the face of dictators and authoritarian regimes.
Speaker 00:00:20
Josh Lim: You had people using Wikipedia for nefarious, malicious ends.
Speaker 00:00:26
And so we felt a responsibility to stop it.
Speaker 00:00:31
We deserve at least to have as much information as possible so that we can make the most informed decisions.
Speaker 00:00:39
We've gone through so much, right?
Speaker 00:00:40
People have gone through so much, not just here in the Philippines and all other countries too.
Speaker 00:00:44
Heather Ford: The Philippines has gone through a lot.
Speaker 00:00:48
It's a democracy, but it's had its share of authoritarian rule, like the 20 year Marcos dictatorship and more recently, the Duterte Government.
Speaker 00:01:01
Josh Lim: There's a reason why the Guinness Book of World Records says that the greatest theft of a government was Ferdinand and Imelda Marcos.
Speaker 00:01:11
The Marcoses were using Wikipedia to whitewash their atrocities, if you will, because Wikipedia is such a powerful thing, especially among younger people.
Speaker 00:01:20
So they were using Wikipedia to try to reshape their narrative.
Speaker 00:01:24
Then the Duterte's were also using Wikipedia in this manner.
Speaker 00:01:29
We have a responsibility to our readers, as editors of Wikipedia, to tell them the entire truth about these things.
Speaker 00:01:37
The truth speaks to power.
Speaker 00:01:40
Heather Ford: Josh Lim's story is one of thousands.
Speaker 00:01:45
Editors around the world are fighting to keep Wikipedia accurate.
Speaker 00:01:49
Francesca Sidoti: But what happens when the threat isn't someone rewriting history, but nobody writing it at all?
Speaker 00:02:01
Heather Ford: This is All That We touch: Wikipedia.
Speaker 00:02:05
A series about how Wikipedia shapes what you know.
Speaker 00:02:09
I'm your host, Heather Ford.
Speaker 00:02:11
I'm a Professor of Knowledge Technologies at the University of Technology, Sydney.
Speaker 00:02:15
And these are my co-hosts, Francesca Sidoti.
Speaker 00:02:19
Francesca Sidoti: With Heather, I'm part of the wikihistories team.
Speaker 00:02:21
Heather Ford: And Richard Cooke.
Speaker 00:02:22
Richard Cooke: I'm a journalist and author who's just written a book about Wikipedia.
Speaker 00:02:27
Heather Ford: Wikipedia has survived for more than two decades on a simple, radical idea that anyone, anywhere could help build the world's encyclopaedia.
Speaker 00:02:38
But right now, that idea is under pressure.
Speaker 00:02:43
Will it still be written by humans?
Speaker 00:02:47
So there's two directions this could go.
Speaker 00:02:51
The one direction that everybody is worried about AI And another direction that's been hiding in plain sight.
Speaker 00:03:04
So, Richard, you wrote a Wired article about Wikipedia, and there's this phrase that you've become well known for.
Speaker 00:03:13
Richard Cooke: Oh, man.
Speaker 00:03:13
People yelling at me when I'm walking down the street all the time.
Speaker 00:03:17
No, it's 'The last best place on the internet'.
Speaker 00:03:21
And it's also now the title of my book.
Speaker 00:03:24
And what I meant by that is that Wikipedia was one of the few things which had continued this kind of moment of promise from the early internet, when a lot of people were feeling that maybe it hadn't panned out so well.
Speaker 00:03:39
People get nostalgic about both the time before the internet and the time when the internet was less garbage.
Speaker 00:03:46
So Wikipedia, for all its flaws, I think has continued that tradition.
Speaker 00:03:51
As someone put it to me, Wikipedia is terrible, but it's still better than the alternatives.
Speaker 00:03:57
I mean, it's my phrase, but it's your experience, right?
Speaker 00:04:00
Yes.
Speaker 00:04:03
So if we go back to when the World Wide Web gets picked up in the late 90s, early 2000, there's this incredible sense of possibility that there's this phenomenal democratising technology.
Speaker 00:04:18
What did that feel like when you were invested in it in that way?
Speaker 00:04:21
Heather Ford: Oh, like enormous possibility, especially living in South Africa, because it felt like that there was so much potential not to access culture, but actually to participate, to share our knowledge and our creativity with the world.
Speaker 00:04:38
Richard Cooke: Yeah.
Speaker 00:04:39
And there's this kind of civic internet, right?
Speaker 00:04:41
This sense that you can have digital institutions that bolster and supplement and in some cases, replace institutions that aren't working properly.
Speaker 00:04:53
And there's a particular moment in time where the Wikimedia Foundation moves from Florida to San Francisco, and they go there thinking that they are the vanguard of a kind of.org boom.
Speaker 00:05:07
Okay, we've had the commercial internet.
Speaker 00:05:09
Now we're going to have the not for profits.
Speaker 00:05:11
Now we're going to have the cultural institutions.
Speaker 00:05:13
Now we're going to have the community.
Speaker 00:05:16
And they very quickly realise nobody else is coming.
Speaker 00:05:19
We're it.
Speaker 00:05:20
We're the last of the dinosaurs.
Speaker 00:05:22
And the dinosaurs are only six years old.
Speaker 00:05:24
Heather Ford: Yeah.
Speaker 00:05:25
You had Yochai Benkler's really seminal book called The Wealth of Networks.
Speaker 00:05:31
And what he was describing in the US, essentially, but also in Europe, where people were gathering together volunteers, creating incredibly high quality goods and services for free in ways that were rivalling goods and services produced by traditional markets.
Speaker 00:05:56
So things like Project Gutenberg, where they were digitising books for free.
Speaker 00:06:01
Wikipedia, where they were creating encyclopaedias for free ride sharing applications.
Speaker 00:06:07
Richard Cooke: Yeah.
Speaker 00:06:08
They run into this proprietary internet, right, where all of this stuff exists, but it becomes commercialised and they almost get knocked off one by one.
Speaker 00:06:17
You know, file sharing is replaced by Spotify.
Speaker 00:06:20
Couchsurfing is replaced by Airbnb.
Speaker 00:06:22
All of the non-commercial ones are replaced by multibillion dollar tech companies, which have a very different attitude to all of this stuff.
Speaker 00:06:31
And what's so fascinating about Wikipedia is that for all its flaws, it's probably the closest thing that we have to what that internet was supposed to look like.
Speaker 00:06:43
Heather Ford: That's why Wikipedia is the 'best place on the internet', maybe right now.
Speaker 00:06:48
But why do you think it's the 'last best place'?
Speaker 00:06:53
Richard Cooke: One of the changes that really took me by surprise.
Speaker 00:06:58
I'd gone through years of writing and research thinking, okay, well, this is the biggest encyclopaedia that's ever existed.
Speaker 00:07:05
Nothing's really been close to it.
Speaker 00:07:08
And I suddenly realised that this was going to be the last human authored encyclopaedia of its kind.
Speaker 00:07:16
Heather Ford: Hmm.
Speaker 00:07:17
Richard Cooke: There was never going to be anything like this again, because AI, generative AI and things like ChatGPT, were going to be used to write things of this length.
Speaker 00:07:29
And already we've seen an AI written encyclopaedia, which is trying to dislodge Wikipedia, in a way, do the same thing that Wikipedia did to encyclopaedias like Britannica, just sort of defeat it by scale.
Speaker 00:07:43
Francesca Sidoti: You mean Grokpedia, the AI powered encyclopaedia started by Elon Musk to supposedly rival Wikipedia?
Speaker 00:07:52
When actually it's been found to promote conspiracy theories and praise Adolf Hitler.
Speaker 00:07:58
Richard Cooke: That's right.
Speaker 00:08:00
There just is not going to be something written by humans alone in this way ever again.
Speaker 00:08:07
And I felt a real sense of loss about that.
Speaker 00:08:09
That what replaced it would be quite soulless by comparison.
Speaker 00:08:14
Francesca Sidoti: I wonder if this is really new territory for Wikipedia.
Speaker 00:08:18
Not so much.
Speaker 00:08:19
The machines arriving at the gates.
Speaker 00:08:21
They've kind of been inside for years.
Speaker 00:08:24
And Wikipedia let them in.
Speaker 00:08:26
Heather Ford: Yes, completely.
Speaker 00:08:27
Because Wikipedia as a community has always been really encouraging of the use of automation.
Speaker 00:08:36
And actually machine learning in particular.
Speaker 00:08:38
Machine learning is what enables Wikipedia to detect automatically whether an edit is likely to be vandalism or not.
Speaker 00:08:52
Francesca Sidoti: So 'vandalism' is when someone will make a Wikipedia edit that's deliberately incorrect; often rude, sometimes quite funny.
Speaker 00:09:01
Heather Ford: And some people say, are absolutely essential to Wikipedia's ongoing survival, were the vandalism bots, right?
Speaker 00:09:12
Francesca Sidoti: They're like the elves making shoes at night.
Speaker 00:09:14
That's how I think about them.
Speaker 00:09:15
Like scurrying around Wikipedia, just doing this stuff that no one wants to do, but is very important.
Speaker 00:09:21
Heather Ford: Super important.
Speaker 00:09:22
Yeah.
Speaker 00:09:23
So there's always been this relationship between humans and machines that has been very cooperative and productive on Wikipedia.
Speaker 00:09:37
So when AI came first, there wasn't an immediate backlash because they thought, well, you know, this is just another tool in our tool set.
Speaker 00:09:47
The question, though, is the extent to which, now that we have AI in the hands of the public, the risk is people who are editing Wikipedia and they're using an LLM like ChatGPT to find out information about the subject, which is not always verified perhaps; contains references that don't exist; editing the article; inserting these statements, these references, and it not getting caught.
Speaker 00:10:23
The risk there is that the quality of Wikipedia articles declines.
Speaker 00:10:28
And especially if it happens in articles and maybe language editions that are not heavily policed by editors.
Speaker 00:10:38
Francesca Sidoti: And it's happened already.
Speaker 00:10:42
Heather Ford: People don't like to talk about it.
Speaker 00:10:43
It's the second biggest Wikipedia behind English, Wikipedia.
Speaker 00:10:50
Josh Lim: Cebuano Wikipedia.
Speaker 00:10:52
It saddens me that it became the butt of all the jokes about Wikipedia in the Philippines.
Speaker 00:10:58
Heather Ford: Wikipedia editor Josh Lim.
Speaker 00:11:00
Josh Lim: So the language is spoken by about 21 million people in the island of Cebu, as well as in some of the neighbouring provinces.
Speaker 00:11:09
Richard Cooke: You know that Cebuano Wikipedia is a love story, right?
Speaker 00:11:12
Heather Ford: It is a love story.
Speaker 00:11:13
Richard Cooke: Yeah.
Speaker 00:11:13
That is this Swedish professor as a kind of tribute to his wife Lorelei, who is Cebuano speaker.
Speaker 00:11:20
Heather Ford: So this is back in 2012.
Speaker 00:11:23
And that professor is Sverker Johansson and his tribute to his wife.
Speaker 00:11:31
His love letter took the form of a bot.
Speaker 00:11:36
A bot is just a computer program that does one job over and over automatically.
Speaker 00:11:43
And can you guess what job Johansson gave his bot?
Speaker 00:11:47
Francesca Sidoti: Was it writing Wikipedia articles?
Speaker 00:11:49
Heather Ford: Yes!
Speaker 00:11:50
Thousands of articles a day in the Cebuano language.
Speaker 00:11:55
The result was 6 million articles.
Speaker 00:12:00
The world's second biggest Wikipedia, written almost entirely by a machine.
Speaker 00:12:07
Francesca Sidoti: It's not written by AI as we know it now.
Speaker 00:12:10
Like this is almost 15 years ago.
Speaker 00:12:12
Heather Ford: Yes, this was a simple script, basically like filling in templates, and it still swallowed a whole Wikipedia.
Speaker 00:12:21
Richard Cooke: Most of Cebuano Wikipedia is millions of articles about like insects.
Speaker 00:12:27
And a lot of these insects are ones which have no name in Cebuano.
Speaker 00:12:33
So Sverker Johansson and his wife actually made up the Cebuano names for a lot of these.
Speaker 00:12:41
Heather Ford: Now that is power.
Speaker 00:12:42
Richard Cooke: Yeah.
Speaker 00:12:43
Yeah, it is.
Speaker 00:12:44
And I said, "Well, you know, you're kind of speaking on behalf of somebody else's culture." And he rather cheekily said that he thought it was a corrective to the number of white men and English speakers editing Wikipedia, because here was this counterweight from a language from the global South.
Speaker 00:13:03
Heather Ford: By another white man.
Speaker 00:13:05
Richard Cooke: Well, I think that he calls himself 'the author of the author' because he doesn't write these.
Speaker 00:13:11
He started this process checking each article before he published it, checking each one for about a minute.
Speaker 00:13:18
Heather Ford: With his wife.
Speaker 00:13:19
Richard Cooke: With his wife, yeah.
Speaker 00:13:20
Some of them.
Speaker 00:13:21
Yeah.
Speaker 00:13:22
So he worked out that if he checked each of the articles that his bot was writing for one minute, it would take 19 years to check them all.
Speaker 00:13:33
So then it was just being they were just being published with no checking, but they had just machine translations of these big special textbooks which itemise every insect in the world.
Speaker 00:13:44
When people talk about AI in Wikipedia and the threat of AI?
Speaker 00:13:48
In my mind, it looks like all Wikipedia's being like Cebuano Wikipedia, that they're not the reflection of a community anymore.
Speaker 00:13:58
They are this big kind of abandoned hall full of computer written articles about insects.
Speaker 00:14:07
Heather Ford: That's one direction.
Speaker 00:14:08
But there's a second possible future for Wikipedia, one that's been hiding in plain sight.
Speaker 00:14:17
Maybe because it's been happening on Wikipedia's own home turf.
Speaker 00:14:22
And now we bring it out of hiding.
Speaker 00:14:29
In the 19th and 20th century, office buildings, banks, department stores had this incredible way of sending information.
Speaker 00:14:40
Francesca Sidoti: Tubes threaded through the walls, floor to floor, Department to department.
Speaker 00:14:47
A whole hidden network, like a nervous system running through the building.
Speaker 00:14:54
Heather Ford: You'd write your note, roll it up, seal it inside a little brass canister, and drop it into the tube.
Speaker 00:15:03
Francesca Sidoti: A hiss of compressed air.
Speaker 00:15:05
And it was gone.
Speaker 00:15:07
Shooting through the walls, around corners, up through the ceiling, popping out on someone's desk three floors away, seconds later.
Speaker 00:15:16
Heather Ford: I'd like to introduce Michael Falk back as a guest.
Speaker 00:15:20
Michael is a chief investigator on the wikihistories Project and a Senior Lecturer in Digital Studies at the University of Melbourne, and he joined us in episode one.
Speaker 00:15:30
And Michael, you think this system of tubes is actually a pretty good way to describe something being built right now by Wikipedia?
Speaker 00:15:39
Michael Falk: Hmm.
Speaker 00:15:40
One of the dreams of the Wikimedia Foundation and some of the people who work for it, is that all the world's facts should be available for everyone.
Speaker 00:15:51
In this dream, each fact is like a little package, which can be popped into one of those tubes and shot off to whoever needs it anywhere on earth.
Speaker 00:16:01
Francesca Sidoti: Okay, so where do the facts come from before they're packaged into that tube?
Speaker 00:16:07
Michael Falk: The main source of facts for this system is Wikipedia itself.
Speaker 00:16:11
If you need to know the date of birth of a celebrity or the land area of a city, Wikipedia is the best place to look for that.
Speaker 00:16:18
The system is very largely automated, though.
Speaker 00:16:20
There is a big human component, so someone might write some software that trawls through a whole lot of Wikipedia pages and pulls out relevant facts.
Speaker 00:16:29
Francesca Sidoti: Okay, so when you say they go to a Wikipedia page, like what does that look like?
Speaker 00:16:34
If we go to the Wikipedia page of Ned Kelly, for example.
Speaker 00:16:38
So Ned Kelly is basically Australia's answer to Robin Hood.
Speaker 00:16:42
This kind of revered, contested outlaw of the 1800s.
Speaker 00:16:47
What facts would it be looking for?
Speaker 00:16:50
Michael Falk: The system's mostly designed for pulling out structured data things like place names, dates.
Speaker 00:16:57
So the big things it would get from the Ned Kelly page would be when he died the 11th of November, 1880.
Speaker 00:17:04
His occupation, which is every Australian knows is bushranger and other atomistic.
Speaker 00:17:10
Simple facts.
Speaker 00:17:12
Francesca Sidoti: And where are each of these little tubes going?
Speaker 00:17:15
Are they popping onto someone's desk?
Speaker 00:17:17
Michael Falk: So again, if we imagine the whole Wikipedia ecosystem as a 1950s office with all the different departments connected by pneumatic tubes, the facts are packaged up and shot through the pneumatic tubes to a central office Wikidata, which is different to the other Wikipedias.
Speaker 00:17:37
Instead of being a textual encyclopaedia you can read.
Speaker 00:17:40
It's like a giant spreadsheet where all of the facts are held in a tabular format that a computer can understand.
Speaker 00:17:46
Heather Ford: Okay, so we've got Wikidata, which is like this giant spreadsheet or database, and Wikidata was started by the Wikimedia Foundation and received funding from Google and other players.
Speaker 00:17:59
Essentially, the reason why they started it was they said it was becoming really difficult for every language version to update facts on its language edition.
Speaker 00:18:11
So, for example, when the population of San Francisco changed because the latest census came out...
Speaker 00:18:19
They said instead of changing it on every single language edition, why don't we just have the central place where we keep all of these facts and just be able to change them once?
Speaker 00:18:30
And then auto populate every single language version where you found the population of San Francisco.
Speaker 00:18:38
Michael Falk: Mhm.
Speaker 00:18:40
Heather Ford: So, Fran, you looked at Ned Kelly and how it was represented on English, Wikipedia and then on Wikidata.
Speaker 00:18:49
What do they look like?
Speaker 00:18:50
Francesca Sidoti: Thank you for the opportunity to research Ned Kelly.
Speaker 00:18:55
So if you look at the Wikipedia article, you'll get all of this kind of juicy detail.
Speaker 00:19:01
You get representations in popular culture.
Speaker 00:19:04
Shout out Heath Ledger.
Speaker 00:19:05
Extensive quotes from sources, the nuances of the relationship Ned Kelly had with other people on Wikipedia.
Speaker 00:19:12
Ned Kelly is a story, and there's this sense of being able to get a little bit to his place in the Australian psyche.
Speaker 00:19:20
The portrayals in film and paint tangle of convicts and settlers, squatters and selectors, Irish and English that kind of made him possible and made him enduring.
Speaker 00:19:30
On Wikidata, Ned Kelly is Q 301816.
Speaker 00:19:35
Heather Ford: Nice robot voice, Fran.
Speaker 00:19:36
Francesca Sidoti: Thank you.
Speaker 00:19:38
So we're talking here about data points.
Speaker 00:19:40
Place of birth.
Speaker 00:19:41
Place of death.
Speaker 00:19:43
Associated events.
Speaker 00:19:45
Politics aren't there.
Speaker 00:19:47
The arguments aren't there.
Speaker 00:19:49
So an example of this is the Wikipedia page will say that Ned Kelly's exact birth date is unknown, but it was probably December 1854.
Speaker 00:19:58
But the Wikidata page says his date of birth was December 1854.
Speaker 00:20:04
Heather Ford: Wow.
Speaker 00:20:04
So translating that uncertainty into certainty.
Speaker 00:20:09
Francesca Sidoti: Exactly.
Speaker 00:20:10
Heather Ford: And not only that, the meaning is changed because the meaning is that we don't know.
Speaker 00:20:18
It is maybe December 1854.
Speaker 00:20:21
That's very different from birth date, December 1854.
Speaker 00:20:29
So there's lots of efficiency in being able to represent Ned Kelly's birth and death date across every single language version.
Speaker 00:20:37
But the problem is that there's these uncertainties, disagreements, volatilities in the data that editors are trying to represent using these very nuanced language.
Speaker 00:20:51
But when that's translated automatically, most of the time, that uncertainty, that disagreement that is vital to really understanding the meaning of the thing gets completely written out.
Speaker 00:21:05
And that changes the meaning.
Speaker 00:21:11
The more that you do this at scale, the more you rely on this data.
Speaker 00:21:16
You are going to introduce inaccuracies, misunderstandings at scale.
Speaker 00:21:23
Michael Falk: Mhm.
Speaker 00:21:24
There's an obvious limitation to Wikidata, which is that it's only appropriate for machines.
Speaker 00:21:29
Humans don't like reading big tables of facts.
Speaker 00:21:32
They like reading narratives and stories.
Speaker 00:21:34
If you wanted to get Ned Kelly's death date from Wikidata, you could go to Wikidata and see it in a table.
Speaker 00:21:42
So the creator of Wikidata, Denny Vrandecic, has begun a new project to make Wikidata work for humans: Abstract Wikipedia.
Speaker 00:21:51
What Abstract Wikipedia will do is allow someone to write a sentence about Ned Kelly in a strange made up language, which automatically includes his death date in the sentence, and then that sentence can be converted into whatever language you want, whether it's Swahili or Spanish.
Speaker 00:22:09
So if there's one easily translatable version of the Ned Kelly story in Wikidata, then any language edition that wants to add an article about Ned Kelly can now get one automatically from abstract Wikipedia.
Speaker 00:22:21
Instead of having to wait for a human author to write an article in that language about this Australian bushranger.
Speaker 00:22:28
You'd be able to create a huge shared world of facts where everyone's equally represented.
Speaker 00:22:35
So there is this very familiar utopian dream.
Speaker 00:22:40
Richard Cooke: This is a really old idea about having a final version of all the knowledge in the world and being able to reconcile arguments through this.
Speaker 00:22:51
So this is a strain in philosophy about, you know, Esperanto, the language which is supposed to unify all languages, is part of this.
Speaker 00:23:00
Leibniz is part of this, where he thinks that if you have a perfectly logical language, then there'll be no war anymore because people can't disagree about material things.
Speaker 00:23:11
It's utopian, but it's also inspirational.
Speaker 00:23:15
Michael Falk: Absolutely.
Speaker 00:23:16
Heather Ford: Do you think that Wikidata and its creation, Abstract Wikipedia, is a real threat to replacing language wikipedias?
Speaker 00:23:25
Michael...
Speaker 00:23:26
Michael Falk: Yeah, so there's different possible futures for Abstract Wikipedia.
Speaker 00:23:30
The inventors imagine a mixed situation where some encyclopaedias, like English will be mostly written in English, but encyclopaedias for minority languages might rely a lot on abstract Wikipedia to generate most of the content for their encyclopaedia.
Speaker 00:23:46
So is Abstract Wikipedia a threat to the ordinary Wikipedias written in human language?
Speaker 00:23:51
I would give two answers.
Speaker 00:23:53
Number one, it's not a threat because I'm fairly sure it will fail.
Speaker 00:23:57
For me, as a scholar of 18th century literature in the history of computing.
Speaker 00:24:02
It's not possible to come up with a made up language that can easily express a sentence in any of the world's languages.
Speaker 00:24:10
Number two, if it succeeded, I think the designers have thought enough about how to make it work with existing Wikipedias.
Speaker 00:24:17
So basically, each language edition, the editors of that edition will have the right to choose how much Abstract Wikipedia content they include in their language edition, and they can choose to include none if they like.
Speaker 00:24:31
If there's something that really does turn me off about Abstract Wikipedia, it is that, despite its talk of incorporating all the world's languages and incorporating all the world's peoples, it does have this homogenising push that it wants to put everyone into one world.
Speaker 00:24:46
And that's probably the ultimate contradiction of the project.
Speaker 00:24:50
I think that's a pity, because it's not really in the actual spirit of what Wikipedia has actually become in the hands of its editors.
Speaker 00:24:57
Heather Ford: Mhm.
Speaker 00:24:58
It envisions a world where meaning is centralised and that you can just push meaning out into different language versions.
Speaker 00:25:07
And I think that is a very different idea of Wikipedia that we started with.
Speaker 00:25:14
Francesca Sidoti: All of a sudden I started thinking about Lord of the Rings and thinking, it's like the Mordor versus Shire question, right?
Speaker 00:25:21
It's like, do you want one rule to rule them all?
Speaker 00:25:25
Do you want Mordor to rule the whole of Middle Earth, or do you want a Wikipedia that reflects the Shire and then reflects the other parts of middle earth?
Speaker 00:25:35
I think there's always been a tension in Wikipedia about how you represent the local and specific, and the truths that people know from their lives, versus how you represent this universal, abstract, numerical data thing that can efficiently move.
Speaker 00:25:53
And that's always been a tension.
Speaker 00:25:54
So Wikipedia is looking down the barrel of these two alternative futures, and whether those two futures can live alongside each other within the platform.
Speaker 00:26:04
Heather Ford: So knowing everything we've talked about today about AI and the centralisation of facts, all of it, are you more hopeful about Wikipedia's future or less?
Speaker 00:26:16
Richard Cooke: Well, in a way, it's held up better than something like The Washington Post has.
Speaker 00:26:20
And The Washington Post had a 100 year head start on Wikipedia.
Speaker 00:26:24
It's not just Wikipedia that's in trouble.
Speaker 00:26:26
It's all the ways that we make knowledge.
Speaker 00:26:29
Heather Ford: Hmm.
Speaker 00:26:30
Michael Falk: I'm very hopeful about Wikipedia.
Speaker 00:26:32
In the age of AI, the Wikipedian community are extremely passionate and committed.
Speaker 00:26:39
In fact, we all need to learn from Wikipedia how to work with AI because they've been doing this for decades and have a brilliant system for approving, regulating, and controlling bots on the platform, which resulted in a great magnification of human effort.
Speaker 00:26:55
As bots work with humans to make the world's largest encyclopaedia function.
Speaker 00:26:59
Heather Ford: What do you think, Fran?
Speaker 00:27:01
Francesca Sidoti: I don't think I would have become a social scientist if I didn't have quite strong beliefs in the messy and complicated.
Speaker 00:27:09
Mhm.
Speaker 00:27:10
I find it hard to believe that we will end up in a situation where this is enough for people.
Speaker 00:27:17
I think there's going to be a an equal move towards complexity and nuance and kind of humanness.
Speaker 00:27:26
Heather Ford: Mhm.
Speaker 00:27:28
And I think in the end, it really does come down to the demand question, which is, to what extent will people really miss the humans that curate knowledge in the careful way that Wikipedia does generally?
Speaker 00:27:44
Now, Wikipedia says that this will actually be in demand because there is so much material that is of low quality that people will increasingly look to Wikipedia as a reliable source.
Speaker 00:28:01
Richard Cooke: The death of Wikipedia is exaggerated.
Speaker 00:28:05
Yes, it has less traffic than it used to, but so does the rest of the internet.
Speaker 00:28:09
And in proportion, it's going to continue for the foreseeable future.
Speaker 00:28:17
Heather Ford: This is All That We Touch: Wikipedia.
Speaker 00:28:21
Francesca Sidoti: If this episode has taught us anything, it's that Wikipedia doesn't come with easy answers.
Speaker 00:28:26
Ask about its future, and the honest response is "It's complicated".
Speaker 00:28:32
So next episode we put that to the test.
Speaker 00:28:36
Five true or false questions about Wikipedia.
Speaker 00:28:39
The kind that sounds simple until you actually try to answer them.
Speaker 00:28:43
Join Heather Ford with new guest Simon Knight.
Speaker 00:28:46
Play along at home as they go head to head in the Wikipedia quiz.
Speaker 00:28:51
Heather Ford: If you've enjoyed this podcast, follow All That We Touch Wikipedia so you don't miss an episode.
Speaker 00:29:00
I'm Heather Ford and I'm the executive producer of All That We Touch Wikipedia.
Speaker 00:29:05
Francesca Sidoti is the associate producer and Jane Curtis is the senior producer.
Speaker 00:29:12
Sound design by John Jacobs.
Speaker 00:29:16
It's made on Gadigal Country in Sydney, Australia by UTS Impact Studios.
Speaker 00:29:22
Executive producer of Impact Studios is Sarah Gilbert.
