New accessibility pilot on bioRxiv pilot: AI-generated paper summaries at 3 levels (general -> expert). Click Automated Services in the dashboard to view these. We welcome feedback. connect.biorxiv.org/news/2023/11...
I like the idea. But on our recent preprint it is does not perform well at all. For me it looks like it takes key words from the paper and produces sentences. But on the otherhand maybe our text was not that AI friendly. www.biorxiv.org/content/10.1...
Richard--this one for our preprint is a combination on nonsensical and wrong is several statements.... I'm now quite skeptical www.biorxiv.org/content/10.1...
Bad bad bad. Anything like this should be opt-in, but at least have an opt-out option. Doing it without permission is a big no-no. Calling it a pilot is not an excuse.
TBH I see this doing more harm than good. There’s just so much that can go wrong. If you want lay summaries for preprints (& one might reasonably argue these are more appropriate for peer reviewed docs, than preprints) they MUST be appraised by humans who understand the meaning of the paper.
cool, interested to see it in action!
Already learning new things. Had missed this. Great feature 👍
Richard, somewhat unrelated — is there something in the works for an optional graphical abstracts? Or does that exist already? Could make for interesting social media feeds
Awesome. If I've understood correctly, it's for new preprints only. Will it be possible to opt-in for older preprints in the future?
I’ve been wondering where we can request features, specifically, if the bots posting to bluesky could also post each preprint’s keywords, then preprints could get picked up by more feeds, versus just scanning terms in title.