Writing a literature review is thinking. It is reading, choosing, summarizing, cross pollinating ideas and building the foundations for your own research. Every part of that process is necessary to do on your own--offloading any of that onto LLMs is a sign of a crappy academic. - ThreadSky

disabilitystor1.bsky.social • 27 days ago

Writing a literature review is thinking. It is reading, choosing, summarizing, cross pollinating ideas and building the foundations for your own research. Every part of that process is necessary to do on your own--offloading any of that onto LLMs is a sign of a crappy academic.

Reposted from /home/cyhaxztpeu/threadsky.app/post.php on line 479

Deprecated: urlencode(): Passing null to parameter #1 ($string) of type string is deprecated in /home/cyhaxztpeu/threadsky.app/post.php on line 479
">
Warning: Undefined array key "handle" in /home/cyhaxztpeu/threadsky.app/post.php on line 480

Deprecated: htmlspecialchars(): Passing null to parameter #1 ($string) of type string is deprecated in /home/cyhaxztpeu/threadsky.app/post.php on line 480

Comments

marcemallowphd.bsky.social•27 days ago

💯💯💯

connors.cloud•26 days ago

New tools change how the work is performed. What you assume to be most important about the profession isn't necessarily true. This is more being worried about the perception of useful work versus the reality of it, pushing the boundaries of knowledge is not to be gate-keeped.

majeff.bsky.social•26 days ago

Using AI to "write" a literature review isn't pushing the boundaries of knowledge. It's making writers lazy and stupid.

majeff.bsky.social•26 days ago

Hell, if AI can write it, you don't even need to read any of it!

navalairhistory.com•26 days ago

Wrong

sschwinghamer.bsky.social•26 days ago

Chris, I don't mean to be rude, but your arguments don't sound like they're based on education or experience in the field of history...are you a historian?

Your bio is blank, so... 🤷‍♂️

seeadler.bsky.social•26 days ago

The thing most essential to expertise is knowing what you are talking about. Outsourcing that act of learning to LLMs not only disrespects your profession, it also disrespects your own participation in it, by obviating the most necessary component of it.

And it's "gate-kept," Chris.

connors.cloud•26 days ago

You are talking about two different things here.

- The act of learning being replaced by an LLM.
- The act of writing academic papers with the aid of an LLM.

The first one I can understand, the second one I don't understand. If you wrote the paper with the aid of another person it's fine.

navalairhistory.com•26 days ago

An LLM is not a person

seeadler.bsky.social•26 days ago

I am an academic editor. I guarantee you I know more about writing academic papers than you do. You cannot write a decent academic paper with the assistance of an LLM. I have seen hundreds of attempts to do so. LLMs hallucinate, and they are extraordinarily incapable of technical & linguistic nuance

connors.cloud•26 days ago

I don't disagree, but soon you're going to be very surprised when they start to get passed you without knowing. Good luck to you. :)

seeadler.bsky.social•26 days ago

*past

One of us is good at English, Chris, and it isn't you, so I'm going to go right ahead and distrust your literary analyses.

getwired.bsky.social•26 days ago

Gate-kept.

The irony.

barryfreed.bsky.social•26 days ago

I’m giving a library workshop on how to do a literature reviews this week and if you don’t mind I’d like to use your skeet for a ppt slide

disabilitystor1.bsky.social•26 days ago

Sure!

pgmcgonagle.bsky.social•26 days ago

Thanks for exposing some of them.

disabilitystor1.bsky.social•26 days ago

Truly bewildering to me how many people think embracing AI aka our/their own oblivion is the way forward. As if administrators and techbros are interested in anything other than profit/eventual automation.

mark-ungrin.bsky.social•27 days ago

That said, we are where we are because of manipulated literature reviews, and guidance created via mindless box ticking - all by real humans.

There's utility in e.g. requiring medical guidance writers to explain and justify when their guidance differs from AI summaries.

https://bsky.app/profile/mark-ungrin.bsky.social/post/3lgxpdsjs6c2h

mark-ungrin.bsky.social•27 days ago

LLMs now routinely produce demonstrably, objectively superior guidance to that generated by the large majority of public health and infection control experts, for example.

Not that LLMs are amazing.

Just that the bar is really, really low.

Seriously - ask one about aerosol transmission of COVID.

drsmith.bsky.social•27 days ago

True. The bar is really really low.

From Jeffrey Epstein’s father in law setting up the modern academic publishing industry to the announcements from RetractionWatch, Science (tm) is filled with gaming, self-interest and all out corruption.

mark-ungrin.bsky.social•27 days ago

Abso...

https://bsky.app/profile/mark-ungrin.bsky.social/post/3led4mor5j22j

mark-ungrin.bsky.social•27 days ago

...freakin...

https://bsky.app/profile/mark-ungrin.bsky.social/post/3lgtkkqzazc2f

mark-ungrin.bsky.social•27 days ago

...lutely.

https://bsky.app/profile/mark-ungrin.bsky.social/post/3lfc47bxkc22s

rocketsredglare.bsky.social•26 days ago

Exactly this. Also, LLMs are by nature derivative, which is kind of the opposite of pushing the boundaries of knowledge so

majeff.bsky.social•27 days ago

LLMs make people stupid and lazy

kikicurious.bsky.social•26 days ago

AND evaluating the quality of the research. I sometimes wonder if a review is bad by how uncritically / credulously they cite.

jowiph.bsky.social•27 days ago

“Presumably scholars check their work” yeah, buddy, the people that use ChatGPT to do their work for them are paragons of academic integrity

thelasttheorist.bsky.social•27 days ago

If someone can't or won't summarize the research they're about to talk about (and presumably add to) I don't really give a shit what else they have to say. Never mind that the AI summary will have errors and imagine things.

grumpycarpenter.bsky.social•25 days ago

And it begs the question "Why even do this in the first place?" The interesting and important synthesis you're describing is occurring INSIDE A HUMAN BRAIN, and to offload this to software seems to obviate the entire purpose of the exercise.

shellinthesky.bsky.social•27 days ago

I note this rationale applies to getting research assistants to write your lit review for you, which of course tenured academics have been doing for ever.

sharpshard.bsky.social•27 days ago

I voiced these concerns to my professor last year as the uni caved to student use of llm. I'm a mature aged undergrad. Still astounds me that I made the argument (& they didn't) & they capitulated to ai use as long as students declared where & it was used. Grim times for thinkers & proto thinkers.

cabbagesandkings.bsky.social•26 days ago

And also scabbing

drlindseyfitz.bsky.social•26 days ago

Can’t believe this even needs to be stated.

sirlolsworthy.bsky.social•27 days ago

'Presumably scholars check their work' LMAO ok bud. I feel that using it for something like, a listing of less well known but still very relevant papers you might otherwise miss is ok but if you think using chatGPT to 'summarise' research & consensus on a topic is fine, give your degree back

sirlolsworthy.bsky.social•27 days ago

one of my most strongly held beliefs is that ALL kids need to be taught how to do lit reviews in highschool. They need to know what cherry picked data looks like, asking leading questions, using low quality sources/refs, definitive statements w/ no real evidence, how stats do/don't work etc etc

sirlolsworthy.bsky.social•27 days ago

They also need to learn that anyone who can't or won't provide sources for claims, instead telling them to do their own research 'like the rest of us had to', cannot be trusted. If it's so important why not share where it came from? Bc they need to feel powerful, 'I know something you don't, SHEEP!'

clbergpowers.bsky.social•27 days ago

Writing the literature review is a critical part of building the knowledge to do the next part. I just... We're not saying it because that's the whole thing.

jackieseidel.bsky.social•27 days ago

👏!!!!
It is THINKING.
That’s what I keep telling students and colleagues too when they say AI helps them so much. Thinking is hard and necessary work. And sometimes joyful. If you think a new thought! And write it down.

drnadadur.bsky.social•26 days ago

This is such a corporate approach to writing: as if only the end product matters. Whereas, writing is thinking, lit. review is thinking. The process that is so meaningful. Checking if the sources are right alone doesn’t do it.

rheinze.bsky.social•26 days ago

I feel like I’m going crazy. Writing two lit reviews atm it’s not „summarizing“ wtf!

disabilitystor1.bsky.social•26 days ago

Im writing my book, and there are parts of it that include lit reviews (and Im trying to smooth out so it doesnt look like that) but ffs, it is NOT summarizing, its thinking!!

rheinze.bsky.social•26 days ago

I find it harder than writing about my own research. It involves a lot of engaging with other’s arguments, understanding and weighing them first and then holding them up against other texts and arguments, it’s a lot of chaos and structuring the chaos and sitting in others thoughts

tkinias.bsky.social•26 days ago

oh yeah, it can be the toughest part to write for sure!

khayyat.bsky.social•26 days ago

Exactly. They're not just summaries. They draw connections between ideas, establish a genaeology of a discipline, judge relative merits, highlight pitfalls and speed bumps, let you talk shit about colleagues, etc.

The LLM can't understand WHY it is doing the lit review, so what can it add?

erinbartram.bsky.social•27 days ago

All of this feels like when you let students bring a sheet of notes into the exam and one of them pays someone else to write it for them and then they don't understand why they don't ace the test.

Comments

Posting Rules

Reply