Lightnews — Scholar-powered news

Dr. Casey Fiesler @cfiesler.bsky.social · 21h

I have not had this experience!

1 1

Dr. Casey Fiesler @cfiesler.bsky.social · 21h

the ick I get at Sam Altman describing Sora as fanfiction... copyrightlately.com/openai-backt...

“We are hearing from a lot of rightsholders who are very excited for this new kind of ‘interactive fan fiction’ and think this new kind of engagement will accrue a lot of value to them, but want the ability to specify how their characters can be used (including not at all).

3 7 37

Dr. Casey Fiesler @cfiesler.bsky.social · 21h

Sometimes I imagine what it might be like to teach an entire course on technology and intellectual property instead of just a single class. I feel like the students get the most absolutely chaotic but hopefully interesting brain dump from me in that class though haha.

1 21

Dr. Casey Fiesler @cfiesler.bsky.social · 1d

If you click through a subtle link on results you do get to a decent disclaimer/warning from Westlaw, but wouldn't it be a good idea to FORCE users to see this (and to click "I understand") before using the tool? And also ideally an educational explanation for how it works and why it can be wrong?

screenshot of "AI-Assisted Research" that includes a small link labeled "How the AI works" above the question

How AI-Assisted Research works
Close
AI-Assisted Research uses large language models - a type of generative AI - and focuses the models on the language of cases, statutes, and other primary law to improve accuracy.

In addition, primary law is referenced in the responses with the actual language from the source, and links are included to read the full primary law documents. Even with these and other precautions, AI-Assisted Research can occasionally produce inaccuracies, so it should always be used as part of a research process in connection with additional research to fully understand the nuance of the issues and further improve accuracy.

The AI-generated summary of results above the list of primary law authority can be extraordinarily useful for getting an overview of the issues and pointers to primary authority, but it should never be used to advise a client, write a brief or motion for a court, or otherwise be relied on without doing further research.

Use it to accelerate thorough research. Don't use it as a replacement for thorough research.

1 5

Dr. Casey Fiesler @cfiesler.bsky.social · 1d

But importantly, Westlaw makes no attempt to explain WHY it generates fictitious sources. There is zero attempt at education here, which I think is usually the case even when LLMs have disclaimers.

And again, I highly suspect that legal technology companies are downplaying limitations.

2 15

Dr. Casey Fiesler @cfiesler.bsky.social · 1d

I just had a look at the Westlaw AI tool, blamed by one of the lawyers in the article (re: them not understanding at the time that it used AI and could generate fictitious sources). This is the disclaimer on the main page for the tool, but there's nothing included with generated results.

AI-Assisted Research uses generative AI and can occasionally produce inaccuracies, so it should always be used as part of a research process in connection with additional research where primary sources are checked to fully understand the nuance of the issues and further improve accuracy.

1 9

Dr. Casey Fiesler @cfiesler.bsky.social · 1d

... I suspect even when not mentioned there is overtrust in AI here. As I mentioned in my original thread I think it's really important that lawyers are educated about the limitations of AI and sufficiently scared of hallucinations.

But this issue re: legal research AI tools is really concerning!

2 14

Dr. Casey Fiesler @cfiesler.bsky.social · 1d

In a great piece of work @404media.co (via @jasonkoebler.bsky.social ) analyzed court records "where a lawyer offered a formal explanation or apology" for problematic AI use: www.404media.co/18-lawyers-c... There is more blame on overwork than lack of knowledge, though...

Dr. Casey Fiesler @cfiesler.bsky.social · Aug 12

A lawyer in my social media comments is telling me that it's "cruel" to suggest that lawyers should be ethically accountable for mistakes introduced by AI because the weight of technology's flaws shouldn't be on burned out lawyers.

And like, all sympathy to junior associates, but also...

2 7 28

Dr. Casey Fiesler @cfiesler.bsky.social · 2d

Not even so much about the change, but the *timing* of announcing was borderline cruel for:

(1) Students who assumed they were eligible and have already been preparing materials
(2) Students who assumed they could apply next year and have to scramble to apply now

www.science.org/content/arti...

‘Completely shattered.’ Changes to NSF’s graduate student fellowship spur outcry

The announcement comes months later than usual, leaving many would-be applicants stranded

www.science.org

4 14

Dr. Casey Fiesler @cfiesler.bsky.social · 3d

An informal poll for fellow academics: What is an appropriate/typical range for numbers of papers to review on an annual basis? Either pure numbers or e.g. proportionate to the number of papers that you submit per year. I'm curious what folks' heuristics are for this.

5 1 3

Dr. Casey Fiesler @cfiesler.bsky.social · 4d

I'm working on some new standup material, and I have this joke about how thanks to my sadistic constitutional law professor who used cold calling and the socratic method I know way too much about constitutional law, and now every day since January has been a bad day.

2 3 61

Dr. Casey Fiesler @cfiesler.bsky.social · 4d

Oh yeah to be clear this wasn't about FERPA. The question was about "demanding" that their own child show them their grades even if they expressed that they didn't want to.

1

Dr. Casey Fiesler @cfiesler.bsky.social · 4d

Ugh.

1 1

Dr. Casey Fiesler @cfiesler.bsky.social · 4d

I made the mistake on commenting on a random video I saw where someone was asking for opinions.

Anyway my opinion is that if parents choose to help their child pay for college, they are still not *entitled* to information about their *adult* child's grades. Apparently this is an unpopular opinion.

5 4 59

Dr. Casey Fiesler @cfiesler.bsky.social · 8d

floating a thought for feedback:
if the liar's dividend is the benefit bad actors can receive from a world in which there is so much doubt about what is real and what isn't,
I was thinking about "the librarian's dividend" re: the value of the people and institutions who help us evaluate information

1 7 20

Dr. Casey Fiesler @cfiesler.bsky.social · 12d

I’m spending today at a big staff development event (few hundred people) for Arapahoe Libraries, focused on AI. I’m running sessions about ethics. Am really eager to get a sense of the vibe and what kinds of questions people have… (I also just love hanging out with librarians.)

1 15

Dr. Casey Fiesler @cfiesler.bsky.social · 18d

At any given time I am usually reading one audiobook, one physical book, and one book on my kindle. This is a pretty good representation of the variety of my tastes at the moment. :)

Screenshot from story graph. Current Reads: The Thursday Murder Club, Katabasis, The One.

2 43

Dr. Casey Fiesler @cfiesler.bsky.social · 20d

I suspect the poster is probably asking about some specific form of AI. (Though if not I guess I’m pretty fond of the predictive algorithm on my insulin pump that’s helping to keep me alive.)

Though regardless I think people would just give very different answers to this question.

1 2

Dr. Casey Fiesler @cfiesler.bsky.social · 25d

Cold calling a student and then grilling them.

Dr. Casey Fiesler @cfiesler.bsky.social · 25d

Ok I had literally not thought of writing off my standup classes on my taxes until this moment.

6

Dr. Casey Fiesler @cfiesler.bsky.social · 25d

I had the random realization the other day that cold calling with the socratic method is basically the same as comedians doing crowd work.

5 2 36

Dr. Casey Fiesler @cfiesler.bsky.social · 28d

This is a really interesting article about theologians' reactions to sensationalized AI-generated videos of Bible stories. www.npr.org/2025/09/07/n...

Some of the debate here re: the power of the original text makes me think of LLMs turning complex scholarly works into pithy podcasts.

Fantasy or faith? One company's AI-generated Bible content stirs controversy

"The AI Bible is a way to really bring these stories to life in a way that people have never seen before. Think of if we were like, the Marvel Universe of faith," said one of the site's creators.

www.npr.org

2 10

Dr. Casey Fiesler @cfiesler.bsky.social · Sep 7

Okay, confession. I did actually start writing this as a cozy(ish) mystery a few years ago, though only about a chapter. :) (Also should I start it up again now, this first scene would have to be quite different since I assume that the conversational robots would just be running on ChatGPT. 😭 )

“Hello.”
“Hello.”
“Goodbye.”
“Goodbye.”
“Hello.”
“Hello.”
It was unclear how long the conversation between two small, gray, vaguely mouse-shaped robots had been stuck in this endless loop. Every pattern needs a stop condition after all, and this loop was a sign of an interruption in the coding process—because these robots’ creator was not the sort to have forgotten a stop condition. This was not an undergraduate in an introductory programming class; this was a tenured professor of computer science at comparable-to-an-ivy major state university.
Was, quite literally. Twenty minutes ago he had been on the path to promotion to a full professorship, and (though he hadn’t yet known this) on the short list for a MacArthur Fellowship. Now, however, he was on the path to his name carved into a plaque on a memorial bench.
It would not be unusual for him to be sleeping in his office, though typically it would be on the couch. But regardless of his position, it was hard to ignore the pool of dark red blood staining the cement gray carpet around his gray-peppered black hair.
“Hello?” This time the word came from outside the door, followed quickly by a soft knock. “Professor Singh? Class started twenty minutes ago and the other teaching assistants are doing some Q and A about the upcoming final exam, but we thought the schedule had a lecture from you on adversarial machine leaning in the—“
As the small, ginger-haired woman pushed open the door she stopped in mid sentence. She didn’t scream, just stood there staring in disbelief at the dead body, as if it were a bug in her code.
“Hello,” said one of the robots.

2 1 13

Dr. Casey Fiesler @cfiesler.bsky.social · Sep 7

So when it comes to fiction skewering academia (re: Katabasis) I feel obligated to remind you that I once tweeted the plot for an entire first season of an academic murder mystery.

Sometimes I think I should write this as a cozy(ish) mystery novel. :)

www.tumblr.com/cfiesler/187...

Post by @cfiesler · 3 images

💬 1 🔁 69 ❤️ 153 · the tenure-track detective agency · I tweeted about a dream, then realized it should be a television show, so I tweeted the whole first season plot. Featuring an academic who ha…

www.tumblr.com

4 2 57

Dr. Casey Fiesler @cfiesler.bsky.social · Sep 6

“Fewer people die in industry.” omg this book is killing me

Besides no one really meant it when they said alt academia was just prestigious or more commonly that there was no shame in it really they meant it even less when they emphasize that all academia paid better had kinder hours was less stressful. Gave you better job security made you happier oh magicians do really well and consulting they said employers like critical thinking and problem-solving skills. They said if fewer people die in industry, they said these offers were uttered by 10 your professors who had already caught the golden goose who could comfortably know they would never face the terrorist their students now dead oh, he took a job and industry. They would say as if industry here was a euphemism on like a farm for old sick dogs and they said it was a kind patronizing little that betrayed what was truly meant alt academia meant failure. The life of the mind unfettered from commerce was the only kind worth living

1 2 10