Post

Sten Linnarsson
@sten.linnarsson.org

Of course in retrospect it's obvious but I only realized now: *all* the work on the publisher's side is amenable to automation by AI. It's inevitable. We'll be human scientists working in real labs on one side, and AI on the other. Doesn't even need to be fancy AGI AI, just regular shitty AI will do

Mike Fainzilber@mfainzilber.bsky.social

Words fail me...

1 repost5 likes

Mike Fainzilber
@mfainzilber.bsky.social

I am afraid that there will be AI on both sides, instead of real labs there will be paper mills submitting manuscripts at high speed, for publisher AI’s to process. The real science will most likely be drowned in this garbage tsunami…

Faraz Mardakheh
@mardakheh.bsky.social

The only way I can think of overcoming this is for journals to verify ‘genuine’ from fake in collaboration with academic institutions. This will be the only way to distinguish who is a real author and who is fake, which will mean only trusting certain institutions for access to ‘credible’ journals.

Sten Linnarsson
@sten.linnarsson.org

Receiving manuscripts, communicating with reviewers & authors, making decisions, typesetting, updating the website, billing, advertising... it'll be AI all the way, for all but a small number of truly serious journals. I would be shocked if this is not already happening at MDPI and Frontiers