AI, ChatGPT and WP

You can talk about anything related to Wikipedia criticism here.
Post Reply
User avatar
wexter
Sucks Warrior
Posts: 574
Joined: Sun Nov 15, 2020 4:18 pm
Has thanked: 274 times
Been thanked: 281 times

AI, ChatGPT and WP

Post by wexter » Sat Jan 28, 2023 2:43 pm

I am using Chatgtp to rewrite and proof letters - it can be used to quickly pump out chapters in a book which are organized and readable. It is an incredibly powerful tool for proof reading and paragraph construction.

Wikipedians want to keep Chatgpt generated stuff out because it lacks sourcing and therefore not verifiable

Really Wikipedians just use insanity, mental illness, and the lowest common denominator to stomp on anything they don't like; including retired physicist who is not so retired.

https://en.wikipedia.org/wiki/User:XOR%27easter
Hi, I'm a physicist. I started editing here because I wanted a hobby where my bookish habits could, in some small way, help people. Since at least the start of this year, I've been burned out on this place, because nearly everything comes back to cleaning up after incompetence, laziness, advertising, or ego. I tried to focus my efforts on the more often-visited and conceptually central science articles, in order to feel like I wasn't wasting my time — the disrepair of such pages is downright demoralizing. This ended up not being sustainable either. Whenever I felt the slightest bit of optimism, a behind-the-scenes policy debate would crush it. So, I'm pretty much gone now.
Xoreaster on Signpost
Chatbot "AI" text is vaguely-plausible bullshit, produced by an industry whose primary output is vaguely-plausible bullshit, marketed to launder the idea of vaguely-plausible bullshit as cute and whimsical. Anyone routing that sewage pipe into Wikipedia should be indeffed and, for good measure, forced to drink orange juice after brushing their teeth. XOR'easter (talk) 02:33, 11 December 2022 (UTC)
I am using Chatgtp to rewrite and proof letters - it can be used to pump out chapters in a book.
Wikipedians want to keep Chatgpt generated stuff out because it lacks sourcing and therefore not verifiable

https://theconversation.com/the-chatgpt ... ive-195908

"Improve this letter by making it more professional and easier to read ideally the letter should be written at the twelfth grade level "

It would be quite ironic to use Chatgpt to write stories about Wikipedia and Wikipedians. Chatgpt "thinks" it is a nail in Wikipedia's coffin (it does not know Wikipedia is not an encyclopedia) the overview folks at Wikipedia/WMF see something coming. It is the fastest way to organize information as to make it presentable and easy to understand for most people.


Large language models on Wikipedia: friends or foes? ----For the last few weeks, a discussion at the Village Pump has been ongoing about the potential use of text generated by large language models (like GPT-2, GPT-3, GPT-J and ChatGPT) in Wikipedia; near the end of December, a thread on the wikimedia-l mailing list discussed the issue as well, going over the potential benefits, drawbacks and use cases for these models.

Recent coverage in Slate describes a series of discussions which took place at Talk:Artwork title concerning the article (Artwork title), whose initial draft consisted of prompted output from ChatGPT. Various demonstrations have been done of the use of these models to assist in writing, editing and formatting (User:JPxG/LLM demonstration and later User:Fuzheado/ChatGPT). Currently, a proposed set of guidelines for the use of these models is under discussion (and, frankly, could use some more eyes on it).

The August 2022 issue of The Signpost explored some of these issues as they concern this publication; see the From the editors page titled "Rise of the machines, or something" for an introduction. – J, B

https://en.wikipedia.org/wiki/Wikipedia ... ed_content
write a story as to why a program in the future like chatgpt will likely replace Wikipedia

In the year 2050, the world had become heavily dependent on technology for every aspect of daily life. People were constantly connected to the internet and relied on it for information and communication. One of the most popular sources of information was a program called ChatGPT, which was a highly advanced language model that had been trained on an unprecedented amount of data.

Wikipedia, the online encyclopedia that had been in use for decades, was still around but it was struggling to keep up with the fast-paced world of technology. People found it cumbersome to search for information on Wikipedia, as it required them to navigate through multiple pages and links. In contrast, ChatGPT was able to understand natural language queries and provide accurate and concise information in seconds.

The creators of ChatGPT had also implemented a feature that allowed the program to continuously learn and update its knowledge base. This meant that it was always up to date with the latest information and had a much wider range of knowledge than Wikipedia.

As more and more people turned to ChatGPT for their information needs, Wikipedia's traffic and usage began to decline. The website's volunteers and editors found it increasingly difficult to attract new contributors and keep the site updated. In the end, Wikipedia became a shadow of its former self, with most of its content outdated and inaccurate.

ChatGPT, on the other hand, continued to evolve and improve, becoming an integral part of everyday life for millions of people. It was the go-to source for information and knowledge, and it had effectively replaced Wikipedia as the world's primary source of information.



Regenerate response
Wikipedia - "Barely competent and paranoid. There’s a hell of a combination."

User avatar
ericbarbour
Sucks Admin
Posts: 4592
Joined: Sat Feb 25, 2017 1:56 am
Location: The ass-tral plane
Has thanked: 1141 times
Been thanked: 1831 times

AI, ChatGPT and WP

Post by ericbarbour » Sat Jan 28, 2023 7:07 pm

wexter wrote:
Sat Jan 28, 2023 2:43 pm
Xoreaster on Signpost
Chatbot "AI" text is vaguely-plausible bullshit, produced by an industry whose primary output is vaguely-plausible bullshit, marketed to launder the idea of vaguely-plausible bullshit as cute and whimsical. Anyone routing that sewage pipe into Wikipedia should be indeffed and, for good measure, forced to drink orange juice after brushing their teeth. XOR'easter (talk) 02:33, 11 December 2022 (UTC)
He's not totally wrong; AI software is STILL at a very primitive state of development. It's not a magical problem solver and has to be used carefully. But his response is exactly like the artists on sites like DeviantArt bitching that people using AI generated art can snipe commission money away from their own hand-drawn work. If you think furry and anime-like porn is an "industry of importance", it's an easy cheap shot. Don't complain at me if the rest of the world disagrees.

Copied from another thread:
I should also note the stupid Stephen Harrison item that ran a few days prior....that guy should be writing corporate propaganda for Disney or an oil company. It seems to be his reason for existence. He is openly being fed "talking points" by WP insiders.
https://slate.com/technology/2023/01/ch ... icles.html
I expect the Wiki*Whores will be totally hypocritical about AI generated content. They LOVE their software bots, some of which have been used for 10+ years to add or modify text and data. But ChatGPT is merely a more complex version of their damn bots, and it does not matter. Some of the little bastards want to BANBANBAN its use already.

Don't believe me? Read this giant Village Pump squabble for a good example. Running for almost two months and not slowing down. And soon enough, a complex and nuanced subject was reduced to a succession of smug nerds screaming "SUPPORT BAN". That's Wikipedia for ya: reactionary as hell whenever someone shows up with any changes, good or bad, to the way they "do business".

Already this is one of the best examples of "Not Invented Here" I've ever seen. Most corporations and other outfits that see an outsider trying to introduce a change or a new product, will say no QUIETLY. On Wikipedia they practically erect 200-foot billboards openly displaying their hatred of "outsiders" and general hypocrisy.

User avatar
Ognistysztorm
Sucks Critic
Posts: 372
Joined: Wed Nov 09, 2022 1:39 am
Has thanked: 68 times
Been thanked: 199 times

Re: AI, ChatGPT and WP

Post by Ognistysztorm » Sun Jan 29, 2023 9:00 am

ChatGPT provided the following answer when asked about Wikipedia demise scenario:
It is difficult to say for certain what the exact scenario of a collapse of Wikipedia/Wikimedia would look like, as it would likely depend on the specific reasons for the collapse. However, one possible scenario could involve a loss of funding and support for the organization, leading to a shortage of resources and staff to maintain and update the website. This could result in a decline in the number of active editors and contributors, leading to a decrease in the quality and accuracy of the information on the website. Additionally, the website's infrastructure could become outdated and vulnerable to cyberattacks, further compromising the integrity of the information on the site.

Another possible scenario could involve a loss of trust in the website due to misinformation or controversy surrounding certain articles. This could lead to a decrease in traffic and readership, causing a loss in revenue from advertising and donations. The organization may also face legal challenges and backlash from groups or individuals who feel that their reputation has been damaged by articles on the website.

In either case, a collapse of Wikipedia/Wikimedia would have significant consequences for the dissemination of information and knowledge. It would likely lead to a loss of access to a vast array of information that is currently available on the website, and could also have a negative impact on education and research.

In the event of Wikipedia collapse, it would be likely that other online encyclopedia and knowledge sharing platform would see an increase in traffic and usage but it will be hard to replace the vast amount of information and knowledge that Wikipedia had.

User avatar
wexter
Sucks Warrior
Posts: 574
Joined: Sun Nov 15, 2020 4:18 pm
Has thanked: 274 times
Been thanked: 281 times

Re: AI, ChatGPT and WP

Post by wexter » Tue Mar 21, 2023 11:57 pm

Chat GPT understands and thus can translate Wikipedia gibberish

please provide me with a summary of this discussion which is a complaint;

https://en.wikipedia.org/wiki/Wikipedia ... iWikiWayne

The discussion is about the user WikiWikiWayne (WWW) who was blocked for 48 hours after violating 3RR (three-revert rule) and was warned for self-archiving a discussion thread. After the block was lifted, he continued to engage in heated discussion, leading to a thread on WP:AN (administrators' noticeboard) regarding the block. The consensus was that the block was deserved, but some thought it should have been extended due to his post-block comments. However, WWW continues to escalate the situation by taking ill-advised actions and making comments that are borderline legal threats and personal attacks against anyone who does something he doesn't like. Some editors propose blocking him, but others suggest that blocking him may escalate the situation, and instead, it would be better if he could find an article to edit and let the issue fade away.

https://en.wikipedia.org/wiki/Wikipedia ... cle_Solana:

please provide me with a summary of the following discussion which is about a complaint:

The discussion is about an alleged violation of the 1RR on all cryptocurrency-related articles on Wikipedia. The user, Mikemaccana, claims that another user, MrOllie, violated the 1RR rule by reverting content they added about the number of developers using the Solana blockchain platform. Mikemaccana explains that blockchains themselves are developer platforms and added a note to politely explain why this information was relevant. However, MrOllie reverted the content twice, which Mikemaccana considers a violation of 1RR. They also accuse MrOllie of lack of good faith, etiquette, civility, not open to compromise, not willing to discuss on talk pages, and failure to discuss, all of which violate the General Sanctions on Wikipedia. Acroterion, an administrator, joins the conversation and informs Mikemaccana that MrOllie's actions do not constitute vandalism and that this discussion is not to resolve disputes over content. Another user advises Mikemaccana to step back and avoid retaliatory threads full of nonsense claims.

please rewrite this wikipedia jargon into plain english

MikeMaccana was given a one-week block from using Wikipedia because he made personal attacks and harassed others. His ability to use TPA (the ability to send messages to other users through Wikipedia) was also taken away by Acroterion. In addition, he is no longer allowed to contribute to the pages related to Blockchain and cryptocurrencies on Wikipedia.
Wikipedia - "Barely competent and paranoid. There’s a hell of a combination."

User avatar
ericbarbour
Sucks Admin
Posts: 4592
Joined: Sat Feb 25, 2017 1:56 am
Location: The ass-tral plane
Has thanked: 1141 times
Been thanked: 1831 times

Re: AI, ChatGPT and WP

Post by ericbarbour » Wed Mar 22, 2023 1:30 am

wexter wrote:
Tue Mar 21, 2023 11:57 pm
please rewrite this wikipedia jargon into plain english
Ha ha sometimes you can be a real funny guy.
Wait, you ARE joking, right?.....

User avatar
ericbarbour
Sucks Admin
Posts: 4592
Joined: Sat Feb 25, 2017 1:56 am
Location: The ass-tral plane
Has thanked: 1141 times
Been thanked: 1831 times

Re: AI, ChatGPT and WP

Post by ericbarbour » Wed Mar 22, 2023 1:45 am


User avatar
wexter
Sucks Warrior
Posts: 574
Joined: Sun Nov 15, 2020 4:18 pm
Has thanked: 274 times
Been thanked: 281 times

Re: AI, ChatGPT and WP

Post by wexter » Wed Mar 22, 2023 1:57 am

ericbarbour wrote:
Wed Mar 22, 2023 1:30 am
wexter wrote:
Tue Mar 21, 2023 11:57 pm
please rewrite this wikipedia jargon into plain english
Ha ha sometimes you can be a real funny guy.
Wait, you ARE joking, right?.....
No it actually took the "Brain Power" of "Artificial Intelligence" to trans-mutate Wikipedia Jargon and Gibberish to English, using 1.21 gigawatts of electricity from a bolt of lightning.

It is incredibly hard to "anglicize" Wikipedia Jargon and Gibberish...

This is how ChatGPT summarized VolunteerMearek's main statement
Icewhiz is a banned former Wikipedia editor who harassed Volunteer Marek with death and rape threats, as well as doxing. Icewhiz attempted to circumvent his ban by using sock puppets, some of which impersonated real-life people. In a recent article, Professors Jan Grabowski and Havi Dreifuss appear to have trusted Icewhiz's false accusations against Volunteer Marek and failed to credit him appropriately, despite his central role in the article's creation.
The article by Grabowski and Klein can be divided into three parts. The first part serves as an introduction and attempts to link the article to a broader social and political context. The second part contains specific criticisms of Wikipedia articles and provides a "black list" of supposedly evil Polish editors responsible for a "Polish conspiracy" on Wikipedia. The last part is almost entirely composed of the author's personal vendettas, grievances, and disputes, with links that are hard to verify.

These sections do not provide any evidence of a connection between the Polish government and Wikipedia editors, and many of the disputes are initiated by Icewhiz's sock puppets or remaining supporters on Wikipedia.
Or summary of the Trust and Safety statement "Summarize this letter" was the only instruction;
The Lead Manager of Trust and Safety at the Wikimedia Foundation wrote a letter to express his appreciation for the contributions of Dr. Konieczny, a sociology professor at Hanyang University, to the Wikimedia movement. Dr. Konieczny has been a long-time volunteer contributor to Wikipedia and Wikimedia Commons, and has made significant contributions in economics, history, politics, and religion. However, Dr. Konieczny has recently been the target of severe harassment both online and in real life, likely from a problematic volunteer editor named Icewhiz. The letter highlights the risks and personal sacrifices that volunteers like Dr. Konieczny undertake in their efforts to make knowledge freely available to everyone in a neutral way. The letter encourages anyone interested in learning more about Dr. Konieczny's important work to contact the Lead Manager of Trust and Safety.
Total defection of taking responsibility for a failure in Wikipedia's editing processes | Circle the Wagons...
Wikipedia - "Barely competent and paranoid. There’s a hell of a combination."

User avatar
oranges33
Sucks Fan
Posts: 181
Joined: Mon Jan 03, 2022 5:33 am
Has thanked: 18 times
Been thanked: 78 times

Re: AI, ChatGPT and WP

Post by oranges33 » Wed Apr 12, 2023 6:30 pm

Found no way to distinguish btw ChatGPT output and Wikipedia output except the former usually produces less biased output. At the same time, chatGPT still has some Wikipedia bias, because it wasn't trained on *that* much text, and Wikipedia was one of the texts it was trained on. I don't see any way they could reasonably enforce a chatGPT ban, nor any reason they'd want to when push comes to shove.

User avatar
ericbarbour
Sucks Admin
Posts: 4592
Joined: Sat Feb 25, 2017 1:56 am
Location: The ass-tral plane
Has thanked: 1141 times
Been thanked: 1831 times

Re: AI, ChatGPT and WP

Post by ericbarbour » Wed Apr 12, 2023 7:45 pm

wexter wrote:
Wed Mar 22, 2023 1:57 am
This is how ChatGPT summarized VolunteerMearek's main statement
Icewhiz is a banned former Wikipedia editor who harassed Volunteer Marek with death and rape threats, as well as doxing. Icewhiz attempted to circumvent his ban by using sock puppets, some of which impersonated real-life people. In a recent article, Professors Jan Grabowski and Havi Dreifuss appear to have trusted Icewhiz's false accusations against Volunteer Marek and failed to credit him appropriately, despite his central role in the article's creation.
"Artificial intelligence" is still the wrong thing to call ChatGPT. "Artificial comedian" would be more accurate.
Found no way to distinguish btw ChatGPT output and Wikipedia output except the former usually produces less biased output. At the same time, chatGPT still has some Wikipedia bias, because it wasn't trained on *that* much text, and Wikipedia was one of the texts it was trained on. I don't see any way they could reasonably enforce a chatGPT ban, nor any reason they'd want to when push comes to shove.
That's the essential truth: it talks like Wikipedia content becuase it used Wikipedia content. The old IBM "Watson" project did the same thing, and the result: a program that could play Jeopardy! but couldn't save the giant corporation from its final decline. Admit it or not, IBM is slowly dying out, and they did it to themselves.

User avatar
Bbb23sucks
Sucker
Posts: 1345
Joined: Fri Jan 06, 2023 9:08 am
Location: The Astral Plane
Has thanked: 1272 times
Been thanked: 270 times

Re: AI, ChatGPT and WP

Post by Bbb23sucks » Thu Apr 13, 2023 12:45 am

ericbarbour wrote:
Wed Apr 12, 2023 7:45 pm
wexter wrote:
Wed Mar 22, 2023 1:57 am
This is how ChatGPT summarized VolunteerMearek's main statement
Icewhiz is a banned former Wikipedia editor who harassed Volunteer Marek with death and rape threats, as well as doxing. Icewhiz attempted to circumvent his ban by using sock puppets, some of which impersonated real-life people. In a recent article, Professors Jan Grabowski and Havi Dreifuss appear to have trusted Icewhiz's false accusations against Volunteer Marek and failed to credit him appropriately, despite his central role in the article's creation.
"Artificial intelligence" is still the wrong thing to call ChatGPT. "Artificial comedian" would be more accurate.
Found no way to distinguish btw ChatGPT output and Wikipedia output except the former usually produces less biased output. At the same time, chatGPT still has some Wikipedia bias, because it wasn't trained on *that* much text, and Wikipedia was one of the texts it was trained on. I don't see any way they could reasonably enforce a chatGPT ban, nor any reason they'd want to when push comes to shove.
That's the essential truth: it talks like Wikipedia content becuase it used Wikipedia content. The old IBM "Watson" project did the same thing, and the result: a program that could play Jeopardy! but couldn't save the giant corporation from its final decline.
Well, it MAY be possible to distinguish as has been done with images for years now. There is no guarantee, but I can easily see this turning into an arms-race.
ericbarbour wrote:
Wed Apr 12, 2023 7:45 pm
Admit it or not, IBM is slowly dying out, and they did it to themselves.
Maybe they would't be if they hadn't fired all their good developers and replaced them with awful, talentless javascript "developers".
"Globally banned" since September 5, 2023 for exposing harassment.

Post Reply