Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Fraud, so much fraud

science.org

Post reply on HN
Print view View original

41–50 of 868 posts

Re: Fraud, so much fraud

#41

Post by vasco »

neom wrote: post #11

I'm the furthest thing from a scientist unless you count 3,000 hours of PBS spacetime, but I love science and so science/academia fraud to me, feels kinda like the worst kinda fraud you can commit. Financial fraud can cause suicides and ruin in lives, sure, but I feel like academic fraud just sets the whole of humanity back? I also feel that through my life I've (maybe wrongly) placed a great deal of respect and trus…

Pick up an old engineering book at some point, something from mid 1800's or early 1900's and you'll quickly realize that the trust people put on science isn't what it should be. The scientific method works over a long period of time, but to blindly trust a peer review study that just came out, any study, is almost as much faith as religion, specially if you're not a high level researcher in the same field and have spent a good amount of time reading their methodology yourself. If you go to the social sciences then the amount of crock that gets published is incredible.

As a quick example, any book about electricity from the early 1900's will include quite serious sections about the positive effects of electromagnetic radiation (or "EM field therapies"), teach you about different frequencies and modulations for different illnesses and how doctors are applying them. Today these devices are peddled by scammers of the same ilk as the ones that align your shakras with the right stone on your forehead.

Re: Fraud, so much fraud

#42

Post by themanmaran »

brink wrote: post #21
themanmaran wrote: post #13

For all the complaints about AI generated content showing up in scientific journals, I'm exited for the flip side, where an LLM can review massive quantities of scientific publications for inaccuracies/fraud. Ex: Finding when the exact same image appears in multiple publications, but with different captions/conclusions. The evidence in this case came from one individual willing to volunteer hundreds of hours producin…

> For all the complaints about AI generated content showing up in scientific journals, I'm exited for the flip side, where an LLM can review massive quantities of scientific publications for inaccuracies/fraud. How would this work? AI can't even detect AI generated content reliably.

Not in a zero shot approach. But LLMs are more than capable of solving a similar scenario to the one presented:

- Parse all papers you want to audit

- Extract images (non AI)

- Diff images (non AI)

- Pull captions / related text near each image (LLM)

- For each image > 99% similarity, use LLM to classify if conclusions are different (i.e. highly_similar, similar, highly_dissimilar).

Then aggregate the results. It wouldn't prove fraud, but could definitely highlight areas for review. i.e. "This chart was used in 5 different papers with dissimilar conclusions"

Re: Fraud, so much fraud

#44

Post by eig »

This sort of behavior is only going to worsen in the coming decades as academics become more desperate. It's a prisoner's dilemma: if everyone is exaggerating their results you have to as well or you will be fired. It's even more dire for the thousands of visa students.

The situation is similar to the "Market for lemons" in cars: if the market is polluted with lemons (fake papers), you are disincentivized to publish a plum (real results), since no one can tell it's not faked. You are instead incentivized to take a plum straight to industry and not disseminate it at all. Pharma companies are already known to closely guard their most promising data/results.

Similar to the lemon market in cars, I think the only solution is government regulation. In fact, it would be a lot easier than passing lemon laws since most labs already get their funding from the government! Prior retractions should have significant negative impact on grant scores. This would not only incentivize labs, but would also incentivize institutions to hire clean scientists since they have higher grant earning potential.

Re: Fraud, so much fraud

#45

Post by IshKebab »

lkrubner wrote: post #20
themanmaran wrote: post #13

For all the complaints about AI generated content showing up in scientific journals, I'm exited for the flip side, where an LLM can review massive quantities of scientific publications for inaccuracies/fraud. Ex: Finding when the exact same image appears in multiple publications, but with different captions/conclusions. The evidence in this case came from one individual willing to volunteer hundreds of hours producin…

How would that be possible? Novelty is a known weakness of the LLMs and ideally the only things published in peer-reviewed journals are novel.

Detecting images and data that's reused in different places has nothing to do with novelty.

Re: Fraud, so much fraud

#46

Post by mden »

madmask wrote: post #24
neom wrote: post #11

I'm the furthest thing from a scientist unless you count 3,000 hours of PBS spacetime, but I love science and so science/academia fraud to me, feels kinda like the worst kinda fraud you can commit. Financial fraud can cause suicides and ruin in lives, sure, but I feel like academic fraud just sets the whole of humanity back? I also feel that through my life I've (maybe wrongly) placed a great deal of respect and trus…

I agree with you, science fraud is terrible. It pollutes and breaks the scientific method. Enormous resources are wasted, not just by the fraudster but also by all the other well meaning scientists who base their work on that. In my experience no, most fraudsters are not evil people, they just follow the incentives and almost non-existent disincentives. Scientist has become just a job, you find all kinds of people th…

"most fraudsters are not evil people, they just follow the incentives and almost non-existent disincentives"

Maybe I'm too idealistic but why does following incentives with no regard for secondary consequences not evil?

Re: Fraud, so much fraud

#47

Post by jimbokun »

neom wrote: post #11

I'm the furthest thing from a scientist unless you count 3,000 hours of PBS spacetime, but I love science and so science/academia fraud to me, feels kinda like the worst kinda fraud you can commit. Financial fraud can cause suicides and ruin in lives, sure, but I feel like academic fraud just sets the whole of humanity back? I also feel that through my life I've (maybe wrongly) placed a great deal of respect and trus…

I think the error is putting trust in scientists as people, instead of putting trust in science as a methodology. The methodology is designed to rely on trusting a process, not trusting individuals, to arrive at the truth.

I guess it also reinforces the supreme importance of reproducibility. Seems like no research result should be taken seriously until at least one other scientist or group of scientists are able to reproduce the result.

And if the work isn't sufficiently defined to the point of being reproducible, it should be considered a garbage study.

Re: Fraud, so much fraud

#48

Post by LeifCarrotson »

neom wrote: post #11

I'm the furthest thing from a scientist unless you count 3,000 hours of PBS spacetime, but I love science and so science/academia fraud to me, feels kinda like the worst kinda fraud you can commit. Financial fraud can cause suicides and ruin in lives, sure, but I feel like academic fraud just sets the whole of humanity back? I also feel that through my life I've (maybe wrongly) placed a great deal of respect and trus…

You're far from a scientist, so it's easy for you to put scientists/academia on a pedestal.

For most of the people who end up in these scandals, this is just the day job that their various choices and random chance led up to. they're just ordinary humans responding to ordinary incentives in light of whatever consequences and risks they may or may not have considered.

Other careers, like teaching, medicine, and engineering have similar problems.

Re: Fraud, so much fraud

#49

Post by georgeecollins »

CarpaDorada wrote: post #32
slashdave wrote: post #22

Earlier quoted context omitted.

You've given two examples. Please explain why you can extrapolate to all of academia.

From personal experience, it is all I've seen. Could anyone be in a position to extrapolate to all of academia without speaking from personal experience? I'm not speaking of all academics (hence 'most'). It's a statement similar to "Hollywood has a drug problem" or something of that sort. My advice to anyone going into Hollywood would be to stay away from drugs; my advice to anyone going into academia is to treat eve…

I work in Hollywood. I am not sure it has more of a drug problem more than say tech or finance. Maybe it does-- I don't know. The point is when a celebrity is a drug addict you hear about it. When a banker or a lawyer is you don't.

Our experience of things has a lot of bias toward what we want to hear. Generalization plays into sterotypes and ideology.

Re: Fraud, so much fraud

#50

Post by mbreese »

slashdave wrote: post #18
idunnoman1222 wrote: post #6

The amazing part about this to me is that the only reason the authors were caught is image manipulation. The fraud in numbers and text? Not so easy to uncover. Prediction: papers stop using pictures entirely

GenAI will make faking western blots fantastically easy

Many journals now require all versions of a gel image that is used in a figure. So, you’d have to fake the full image that is cropped down to the lanes used in the figure. I think there aren’t as many of those raw images around to train AI on… yet.
Post reply on HN