@PhilipDBunn For one: If you're not a mother-tongue English speaker, don't have access to language services, but have done really great research. and believe that it could be published. It's called leveling the playing fields a bit.
@hiiragi2280 But you put the burden on the other passenger. When I arrive, I have to ask the passenger to move, have to wait for them, look like a jerk while they huff and puff about privileges, & that only other seats are terrible, it's only a few stops... They hope I'm too timid to ask.:-(
@rowbott1 Well, then I guess I'm just more cautious in making a statement. For me, making a statement about a paper, based on its title and authors (no matter how well I know them), and what other people have said about it, is too much of a risk.
I have to disagree. My policy and I teach my students: never cite a text, and never refer to anything in that text, unless you have seen it with your own eyes. NEVER.
Lenka is correct here, Hunter is wrong. This is not a big deal. There are different kinds of citations in a paper on a spectrum from "I actually used this result in detail" (you read those carefully), all the way to "some people did something similar in this space" (you don't really read those in detail because they don't matter much for what you're writing anyway). What is the issue here exactly?
i hate ai detection programs so much. my fully HUMAN WRITTEN essay shows up as 87% ai written. what are you even supposed to do if your teacher brings it up? how can you disprove them?
@mamixusbertius You might not have read the entire paper, but,
1. Make sure the paper exists
2. Make sure the paper actually says what you claim it says.
I realise there is a complication when working with multiple authors. I assume they are doing the same. Maybe I should be more careful there.
@hoofnagle@thartman2u ... of course, once you leave this school & are in the real world, you will have to know how to use AI to conceptualise, outline, draft, revise & translate your work, but we're not going to teach you that. For that, you're on your own. Good luck with trying to crack that.
. Students get low grades
. Admin to teachers: Do something about this!
. Teachers implement active learning, scaffolding, advising, counselling, etc.
. Result: Grades improve
. Students: Yay!
. Teachers: Yay!
. Admin: "Gade Inflation!" Reduce the grades!!!
Sigh:-(
Breaking news: Harvard faculty votes to cap the number of A's awarded in course grades, a big step in combatting the grade inflation that has been dumbing down our courses, conveying the wrong message to students, and making universities a national laughingstock.
Our new study has been published today!
Many #MedEd educators may assume that giving a powerful AI a scoring rubric is enough to evaluate video assignments fairly.
But rubric-only prompting inflated student scores by +26 points. Adding a critical expert-style prompt reversed the bias entirely (-8 points).
We may need prompts calibrated between generosity and criticism. AI video scoring still depends heavily on prompt design, even with models like Gemini 2.5 Pro.
@Nettleshippy I have read it, but I must admit it was because it was part of my course, and once I started, I took it as a personal challenge to complete it. Did I understand it? Now, that is a different question.
@hipidura The publishers have been publishing the articles knowing that this is happening and ignoring it when identified. Now that those errors are called hallucinations, the publishers might take responsibility. The only differences are AI and scale.
Actually, the main reason is that we've been writing and complaining about false citations and false statements about the contents of texts for years. The publishers have done nothing about it. Now, suddenly, AI does it, and they're all prissy. Too little, too late!
I'll tell you why so many people upset about the "no hallucinated citations" ban on the arxiv: because they've all been copying citation lists from each other without checking them since the beginning of time.
And why did they do this? Because half of the citations in scientific papers are politics and not to the benefit of the reader. If you don't list the right papers, your paper doesn't look 'right' and reviewers will complain that you didn't cite this-and-that other unrelated work.
For what I am concerned, these are all bullshit citations that shouldn't be in the papers in the first place. They can easily be automated by "related papers" links, that are (wait for it) provided by... AI...