I've been playing with it on over the last week, and my experience is that that AI is incredibly prone to invent informations that look credible, while being entirely made up by it. It would generate inexistent persons, dates, events and places, and build an entire story out of them that, nevertheless, looks credible. I would advise anyone against using it for anything more than helping with paraphrasing already existing texts, when using it.
Just now. when asked about bishop Pedro Sardinha (first bishop of Brazil) he told me: It's possible that you are referring to Pedro da Silva Sardinha, also known as Father Francisco, who was a Portuguese Jesuit missionary who served as the bishop of Timbuktu in the early 18th century. Father Francisco was born in Lisbon, Portugal, and entered the Jesuit order as a young man. He was sent to West Africa as a missionary and served as the bishop of Timbuktu from 1705 to 1706. Father Francisco is known for his efforts to convert the local population to Christianity and for his work to improve the education and literacy of the people in Timbuktu. However, he died in 1706, apparently while being held captive by local authorities. This person never existed, nor those events, it was entirely made up by the AI, but looks credible. Often it says the information is based on its research on reliable sources, even presenting some of them (but also fake). Yesterday, when asked about the Quintinense Scheme, a well known sockpuppet affair of the Portuguese Wikipedia, it said it was an arrangement by King Manuel I of Portugal in the early 16th century, concerning the way the lands should be cultivated in the colonies, named after Quinta do Lorde, which it said was the first settlement in Madeira (it's actually a tourist resort built in the late 1990s). I'm constantly surprised by the AI capacity of inventing such elaborate fakes. In the case of Pharos experiment, this AI generated information, which still persists (unsourced) in the article, is probably fake as well: The practice of titling artworks dates back to ancient civilizations, where works of art were often inscribed with the name of the artist and the subject of the piece. Best, Paulo Ziko van Dijk <zvand...@gmail.com> escreveu no dia quinta, 22/12/2022 à(s) 18:01: > Hello, > About my experiences with Chat CPT - is it good enough to write our > articles? - I made this video: > https://youtu.be/zKPEyxYt5kg > Kind regards > Ziko > > Am Di., 20. Dez. 2022 um 20:25 Uhr schrieb Pharos > <pharosofalexand...@gmail.com>: > > > > I'd like to share my experiment with using ChapGPT to help write an > article on English Wikipedia: > > > > https://en.wikipedia.org/wiki/Artwork_title > > > > You can see an explanation of the process here, your comments are > welcome: > > > > https://en.wikipedia.org/wiki/Talk:Artwork_title#Use_of_ChatGPT > > > > Thanks, > > Richard > > (User:Pharos) > > > > On Tue, Dec 20, 2022 at 1:45 PM ZhaoFJx <zhao...@gmail.com> wrote: > >> > >> I'd be curious about copyright issues though, as it's licensed on > Github. It may indeed be infringing copyright, since he may require > attribution. However, I agree with The Cunctator that the community won't > do anything about it. > >> > >> Incidentally, I'm a little worried that he might grab broken content > that wasn't rolled back. After all - the current artificial intelligence > can not reach the level of human thinking > >> > >> Sincerely, > >> ZhaoFJx > >> _______________________________________________ > >> Wikimedia-l mailing list -- wikimedia-l@lists.wikimedia.org, > guidelines at: https://meta.wikimedia.org/wiki/Mailing_lists/Guidelines > and https://meta.wikimedia.org/wiki/Wikimedia-l > >> > >> Anders Wennersten <m...@anderswennersten.se> 于2022年12月19日周一 08:52写道： > >>> > >>> I think we should not underestimate what this could evolve into. We > thrived because our readers find us "good enough" when it comes to finding > facts, not the ultimate source. > >>> > >>> And the software learns by reading, and can (and have done so) > Wikipedia, Wikidata etc and represent our data in its own syntax och > present it in a way so it is not a direct copy. Perhaps data will be a bit > delayed to the actual content in Wikipedia, but so what - good enough? > >>> > >>> Anders > >>> > >>> Den 2022-12-19 kl. 14:26, skrev Gnangarra: > >>> > >>> AI simply cant descriminate between good research and faked research, > for any outcome it must provide all of its sources whether they are from > Wikipedia, Wikidata, WikiCommons, WikiSource or some other place. > Otherwise it will answer yes to some asking if the world is flat because > it'll seek out that answer and find all the nonsense that has been produced. > >>> > >>> On Mon, 19 Dec 2022 at 06:02, Erik Moeller <eloque...@gmail.com> > wrote: > >>>> > >>>> On Sun, Dec 11, 2022 at 5:55 AM Anders Wennersten > >>>> <m...@anderswennersten.se> wrote: > >>>> > ChatGPT is now making headlines more or less every day and I > perceive > >>>> > them to try to position themself av the "next" google. > >>>> > >>>> I suspect OpenAI will continue to focus on generative applications > >>>> (images, code, text for purposes such as copywriting, eventually > >>>> music/video) and won't attempt to compete with Google directly, but > >>>> we'll see. Currently GPT-3.5 (which ChatGPT is based on) is very prone > >>>> to generating nonsensical answers, citations to works that don't > >>>> exist, etc. But it is pretty cool if you keep its limitations in > >>>> mind--for example, it's quite good at bootstrapping small scripts in > >>>> various programming languages (with mistakes and idiosyncrasies). > >>>> > >>>> Google has one of the largest AI research programs on the planet, they > >>>> just are extremely conservative about letting anyone try their models > >>>> (due to reputational concerns, e.g., that generative AI will spit out > >>>> racist output within about 30 seconds of people poking its > >>>> guardrails). This blog post from September is instructive about the > >>>> direction they're taking with what's called retrieval-augmented > >>>> generation; see the paper linked from the post for details: > >>>> > >>>> https://www.deepmind.com/blog/building-safer-dialogue-agents > (DeepMind > >>>> is part of Google) > >>>> > >>>> That is likely to yield significantly more accurate answers than what > >>>> ChatGPT is doing, and is difficult to replicate for folks like OpenAI > >>>> without being dependent on the search APIs of big search companies. > >>>> It's worth noting that Google has also started to incorporate language > >>>> model tooling into how it's presenting search results (e.g., > >>>> summarizing or highlighting different parts of a website to make the > >>>> result snippet more useful). > >>>> > >>>> A retrieval-augmented approach that leverages Wikidata could IMO be > >>>> quite powerful and could be a useful research program for Wikimedia to > >>>> pursue, be it independently or in partnership with others. The > >>>> resulting technology should of course be fully open source. > >>>> > >>>> Querying Wikidata via SPARQL is currently still a bit of wizardry (and > >>>> the query builder is extremely limited). To pick a completely random > >>>> example not at all inspired by current events, if I wanted to see a > >>>> list of journalists with Mastodon accounts & a picture, I currently > >>>> have to do this: > >>>> > >>>> SELECT DISTINCT ?personLabel ?mastodonName ?pic > >>>> WHERE { > >>>> ?person wdt:P4033 ?mastodonName ; > >>>> wdt:P106 ?occupation . > >>>> OPTIONAL { ?person wdt:P18 ?pic . } > >>>> ?occupation wdt:P279* wd:Q1930187 . > >>>> SERVICE wikibase:label { > >>>> bd:serviceParam wikibase:language "en" > >>>> } > >>>> } > >>>> > >>>> Make a small mistake (a curly brace missing) and you'll get a red > >>>> error message. Forgot the * after wdt:P279? A different response set > >>>> in ways that are difficult to spot or reason about. > >>>> > >>>> Why can't I type "list of journalists with their picture and Mastodon > >>>> account" as a natural language query? (You can try it in ChatGPT and > >>>> it'll get you started, but it'll generate nonsense P/Q numbers.) If > >>>> such queries could be produced reliably, it could be a very useful > >>>> tool for readers as well. > >>>> > >>>> Warmly, > >>>> Erik > >>>> _______________________________________________ > >>>> Wikimedia-l mailing list -- wikimedia-l@lists.wikimedia.org, > 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/wikimedia-l@lists.wikimedia.org/message/YYTLJVCDSYITUKNA2DJSK5SSR3AZ3B5F/ > >>>> To unsubscribe send an email to wikimedia-l-le...@lists.wikimedia.org > >>> > >>> > >>> > >>> -- > >>> Boodarwun > >>> Gnangarra > >>> 'ngany dabakarn koorliny arn boodjera dardoon ngalang Nyungar > koortaboodjar' > >>> > >>> > >>> _______________________________________________ > >>> Wikimedia-l mailing list -- wikimedia-l@lists.wikimedia.org, > 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/wikimedia-l@lists.wikimedia.org/message/CVXPECMNLGBGIQYP2DI7IRJVLUNNOF6B/ > >>> To unsubscribe send an email to wikimedia-l-le...@lists.wikimedia.org > >>> > >>> _______________________________________________ > >>> Wikimedia-l mailing list -- wikimedia-l@lists.wikimedia.org, > 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/wikimedia-l@lists.wikimedia.org/message/HKAPIBPSXAETLTFQFQDPDCSGCFWDCXAQ/ > >>> To unsubscribe send an email to wikimedia-l-le...@lists.wikimedia.org > >> > >> > >> > >> -- > >> Wikimedia-l mailing list -- wikimedia-l@lists.wikimedia.org, > 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/wikimedia-l@lists.wikimedia.org/message/ALV3QLKTB7RWCCR5W4HNDA4ZDG5ARRDC/ > >> To unsubscribe send an email to wikimedia-l-le...@lists.wikimedia.org > >> > >> _______________________________________________ > >> Wikimedia-l mailing list -- wikimedia-l@lists.wikimedia.org, > 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/wikimedia-l@lists.wikimedia.org/message/ZRVVJ3WVSB2VKRD5MQYILZYHLVZBMSQO/ > >> To unsubscribe send an email to wikimedia-l-le...@lists.wikimedia.org > > > > _______________________________________________ > > Wikimedia-l mailing list -- wikimedia-l@lists.wikimedia.org, 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/wikimedia-l@lists.wikimedia.org/message/UI6THQHUMH7SQLHDA4JUI7FOCXIYXGXI/ > > To unsubscribe send an email to wikimedia-l-le...@lists.wikimedia.org > _______________________________________________ > Wikimedia-l mailing list -- wikimedia-l@lists.wikimedia.org, 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/wikimedia-l@lists.wikimedia.org/message/N5AZFZIZSR5GOBRXCMX4FEBBFGCFBWPJ/ > To unsubscribe send an email to wikimedia-l-le...@lists.wikimedia.org
_______________________________________________ Wikimedia-l mailing list -- wikimedia-l@lists.wikimedia.org, 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/wikimedia-l@lists.wikimedia.org/message/ZA6ZM6HO2TYMKOGFBR6JN2GNXIKI7XCE/ To unsubscribe send an email to wikimedia-l-le...@lists.wikimedia.org
