CⓐmeronNeylon

...and specifically there is evidence of greater use by regions that historically have poor access to closed research outputs. Lots more to dig through and we hope many will, read the paper (and the ~100 pages of supplementary material) for the full story!

What's exciting about this is we've shown at global scale that #OpenAccess leads to more use of research. We tried super hard to identify subsets of the data where this wasn't true but there really aren't many This is robust evidence that #OpenAccess does what it was meant to

Popping back to boost to the great work of @CK_Karl_Huang @COKIproject in Scientometrics Short version: #OpenAccess works! Longer version: Using large scale data we show that #OpenAccess outputs are cited by a greater diversity of actors Longest: doi.org/10.1007/s11192…

Full PhD Scholarship commencing 2024 with @AmyDobsonCU @KkatieElliss at @CurtinUni on "Youth sociality in the era of visual social media: intersectional feminist approaches". Open now; applications due 25/08/2023. Details: scholarships.curtin.edu.au/Scholarship/?i… (Pls RT.)

Check out the latest update from BAD Project Technical Lead, Kathryn Napier: 'The Book Analytics Dashboard Project: Reflections on Coordination, Collaboration, and Community Consultation': openknowledge.community/the-book-analy… Ping @COKIproject @CurtinIC @OAPENbooks @OAeBU_Project

Our @JEPub article about the @COKIproject #OA Dashboard is published! Ping @CameronNeylon @wilsk16 @CK_Karl_Huang @CurtinIC @CCAT_Curtin Article here: journals.publishing.umich.edu/jep/article/id… Checkout the COKI Open Access Dashboard: open.coki.ac #OpenAccess #COKI via @COKIproject

If you are interested in #ethics, #RRI, #biotech #regulation, please get in touch! We have PhD scholarships for domestics and internationals to be based at @UniofAdelaide and @ANU_CPAS Please RT and spread the word.

We are hosting two big public events! One hybrid public debate on “Is #Openscience Good for Research” on March 31, featuring Helen Longino, @CameronNeylon, @ankeny_rachel & @wyatt_sally (registration soon available here: exeter.ac.uk/news/events/de…), and …

@kaveh1000 @dannykay68 @cgreen Terrifying, if true.

@CameronNeylon @dannykay68 @cgreen Thanks. But I was told no one knows more than you do ;-)

QUESTION: I don't know if this is a stupid one. Does LLM in the form of ChatGPT use information in research papers that are behind a paywall? If yes, what is the technology going on there to provide access? If no, is this a huge argument for OA?

@kaveh1000 @dannykay68 @cgreen I don't know - it's a moving target. I think this used to be true and GS ignored structured metadata in favour of their own extraction protocols. This led to some very odd behaviour WRT to repository indexing. But others will know more about the current state than I do!

@CameronNeylon @dannykay68 @cgreen 🤣 OK. A slightly related question. Do you happen to know if metadata in PDFs (Author, title, keywords...) affects indexing by Google Scholar etc? I assumed if we have full XML and HTML then this would have no effect. A client is disagreeing.

@aarontay @dannykay68 @cgreen That's probably true of public releases like GPT-Chat but I'd be staggered if MS and Google (certainly Google) weren't training these models on the internal text they have available (and keeping that internal)

@CameronNeylon @dannykay68 @cgreen I was under impression GPT-3 was basically using common crawl data which might have some full text of OA papers. Unlikely to have paywalled articles. I mostly queried bing+chat but so far I only got info obtainable from title/abstract of paywalled only content.

@kaveh1000 @dannykay68 @cgreen Hardly...more "...if you don't give us the full text we won't index you in our search engine(s)..."

@CameronNeylon @dannykay68 @cgreen How do they have this access Cameron? Paid subscription?

@dannykay68 @cgreen Both Microsoft (OpenAI) and Google have access to full text of the majority of paywalled articles from big publishers. I'd be reasonably confident that they could train on them and by default probably would Not a new argument of course: doi.org/10.1371/journa…

@cgreen Do you don’t know the answer? I hope someone does!

Honoured to be giving the keynote at this year's @ALPSP conference on how responsible research assessment developments may impact on scholarly publishing!

I’m quite troubled that a basic website with very little verified information concerning “predatory publishing” is being shared widely. What about the sanctity of peer review? I don’t think that anyone sharing the link to the site itself can name a single human behind it.

Reminder of next week's OASPA webinar: Reviewing the Data on Knowledge Creation: Access, Governance and Equity - mailchi.mp/oaspa/announci… Join us live on March 7

@jameswilsdon @dannykay68 @COKIproject

Wondering what is going on post-ERA? Puzzled that combining WA unis could possibly increase rankings? We're hosting world-leading expert @jameswilsdon in Perth next week and you can hear him on the future of research evaluation on Wednesday 8th at 5pm eventbrite.com.au/e/public-polic…

Its great to see Australia engaging with international perspectives on research evaluation. We are looking forward to hearing what @jameswilsdon has to say when he visits @CurtinUni as part of his Australian tour! @CameronNeylon