iiinova — where your digital self keeps evolving

Dave Bloom @drumkitdave.bsky.social · 23/08/2026

A big problem in academic databases as well as commercial search engines. Some vendors (looking at you, ebsco) have become so taken with rewarding novices with lots of results for poorly conceived searches that they’ve tanked precison and replicability

170

Dave Bloom @drumkitdave.bsky.social · 10/07/2026

Love this point and would add that EVERY domain has its experts even if they don’t come immediately to mind. We librarians hear “you shouldn’t use ai for xyz, but it’s great if you just need to find some information quickly” regularly from people who haven’t thought at all about information seeking

2315

Dave Bloom @drumkitdave.bsky.social · 10/07/2026

Curious how often this happens in fields or organizations where the messaging around ai stresses that “YOU are responsible for verifying outputs”, isn’t it?

030

Dave Bloom @drumkitdave.bsky.social · 10/06/2026

If you don’t want to do all the following up, you could work with your campus librarians (maybe start with a schol comm librarian). It’s helpful for us to know about this stuff to spot patterns with certain publishers. They may also be able to move things along with the journal or acm in general.

000

Reposted by Dave Bloom

Raina Bloom @mmelibrarian.bsky.social · 22/03/2026

Okay so a few things here, but first and foremost: we have no way to substantiate that these suggestions work, except for the ones that Google describes in their own documentation. And if they’re already in Google’s documentation, then we don’t need this post. But also - 1/

131

Dave Bloom @drumkitdave.bsky.social · 09/02/2026

Well, a wife-abusing cannibal ex-boxer told us we were all fat on behalf of RFK Jr, and Amazon tried to sell us surveillance tech on the backs of missing puppies, so the race to the bottom is competitive, but the AI ads have been more than up to the task.

011

Dave Bloom @drumkitdave.bsky.social · 22/12/2025

Here’s another pretty good one on NotebookLM and summarization. dl.acm.org/doi/pdf/10.1...

dl.acm.org

020

You had the same thought as I did! I explained what I found above.

By November 2023, it included the 2020-2022 run, including the article that Jill linked. Not a sure thing that AI was involved (other citations in this article are sloppy, but not necessarily AI), but, despite claimed publication date, it wasn't on the site until months after ChatGPT came out. end/

030

There aren't Wayback-archived versions of the article or volume/issue, but there's an archived version of the publication archive. Here's the weird part - by September 2022, the publication archive was out of order and didn't have any volumes published after 2020. 4/

120

Anyway, after I did some checking on the article that Jill linked (mixed results--journal isn't in Ulrichs Periodical Directory, but the credited authors are affiliated as described), I used the Internet Archive Wayback Machine to see when this article actually showed up . . . 3/

120

For instance, about a year ago, we found a publisher that populated their archive with a bunch of back-dated abstracts that were almost certainly all AI-generated. It lent them an air of legitimacy presumably so that their more recent, low-quality articles had some weight to them. 2/

122

Academic librarian here and my colleagues and I have done some work on instances like this. A tricky thing about sketchy publications is that you often can't trust dates. This article may have a 2021 data on it, but that doesn't mean it's actually been around since then. 1/

132

I checked through some of the other citations, as well. Most are either legit or only contain minor issues that are pretty common human goofs, like this one (wrong volume/issue, wrong page numbers), but there are some odd things about this article/journal that I'll post further up.

The upshot is that we were able to knock results from the journal out of our article search and got the publisher's journals removed from Ulrich's. But this is clearly going to be an eternal game of Whac-a-Mole for librarians (and for patrons who contact us to report this stuff)! 4/4

2326

In another, a student found an article that had seemingly AI-generated fake citations using our institutional article search (which searches our subscription stuff, but also beyond). It got picked up via CrossRef. Another sketchy publisher, but this one with enough oomph to be listed in Ulrich's. 3/

1183

In one, a researcher told us he'd found himself cited as coauthor on a paper he hadn't written. We found a few years' worth of backlog on the journal's page (hijacked version of a legit journal) was AI-generated abstracts and titles to pad things out for the newer, dubious full-text articles. 2/

1213

Dave Bloom @drumkitdave.bsky.social · 29/04/2025

My colleagues and I (librarians at UW-Madison) just ran a two-hour presentation on the impact of gen AI on predatory publishing for library staff this week! It was directly inspired by a few patron interactions similar to what you're describing. 1/

1286

My bandmates and I were *really* into that show. A few years later, a club offered my ex-bandmates' new band an opening spot for either Flickerstick or the Arcade Fire, who were just about to receive their Pitchfork 10.0 for Funeral. Guess which show my friends picked . . .

Carl T. Bergstrom @carlbergstrom.com · 19/04/2025

Reposted by Dave Bloom

1. The DOJ has sent threatening letters to several medical journals, focused around whether these journals are spreading misinformation that harms the American people and suppressing alternative viewpoints. I study how the norms and institutions of science facilitate discovery and self-correction.

nytimes.com

Trump-Allied Prosecutor Sends Letters to Medical Journals Alleging Bias (Gift Article)

An interim U.S. attorney is demanding information about the selection of research articles and the role of N.I.H. Experts worry this will have a chilling effect on publications.

51938405

Reposted by Dave Bloom

molly the librarian @bigmolly.bsky.social · 06/04/2025

Here's my lightning talk about AI at #acrl2025 this morning. Thanks to @nicoloff.bsky.social for recording. youtu.be/H2UdS1MqH8w?...

youtu.be

Take a Breath. Most Students Don't Need AI Skills

YouTube video by Molly June Roquet

85119

Dave Bloom @drumkitdave.bsky.social · 04/04/2025

In classes where AI is an info lit instruction component, I cover ethics, but focus on pragmatic issues, e.g., databases and search engines are better tools for finding credible sources that you can assess for authority, saving you work & time. It undercuts overall assumptions about AI and ease.

010

Dave Bloom @drumkitdave.bsky.social · 02/04/2025

Could only attend virtually this year, but really enjoying how Ruha Benjamin is complicating what appears to be a lot of "let's throw AI at it" content scheduled for the next few days of the conference. #ACRL2025

0172

Reposted by Dave Bloom

Carl T. Bergstrom @carlbergstrom.com · 17/03/2025

You have to be kidding me. www.nytimes.com/2025/03/17/s...

Even chatbots get the blues. According to a new study, OpenAI’s artificial intelligence tool ChatGPT shows signs of anxiety when its users share “traumatic narratives” about crime, war or car accidents. And when chatbots get stressed out, they are less likely to be useful in therapeutic settings with people.

The bot’s anxiety levels can be brought down, however, with the same mindfulness exercises that have been shown to work on humans.

1031419182

Dave Bloom @drumkitdave.bsky.social · 13/03/2025

Would love a follow up that evaluates sourcing on more typical user queries! Anecdotally, your findings square roughly with what I've noticed about sourcing for more standard information seeking-type prompts, but 'find this exact quote' is pretty distinct from 'answer this and provide support'.

010

Dave Bloom @drumkitdave.bsky.social · 13/03/2025

Although this study would suggest that you might have a difficult time verifying the output for that prompt, because the sourcing may be incorrect, incomplete, or misleading. And if you're not an expert on drugs or genetics, you definitely shouldn't trust the output without verification.

100

Dave Bloom @drumkitdave.bsky.social · 13/03/2025

That phrasing got me at first, too, but the queries in the study were "here's a quote, find me the source," so it mostly just establishes that sourcing is garbage. This is a very big deal, but shouldn't be extrapolated out to reliability of all output.

030

Reposted by Dave Bloom

Raina Bloom @mmelibrarian.bsky.social · 06/03/2025

I had such a good time doing this. Thanks to @alexhanna.bsky.social and @emilymbender.bsky.social for inviting me on!

1174

Dave Bloom @drumkitdave.bsky.social · 28/02/2025

Definitely worrying that an expert in assessment and digital learning is so overwhelmed with the narrative around AI as an efficiency engine that he steps right past a well-established, more reliable solution.

000

Dave Bloom @drumkitdave.bsky.social · 28/02/2025

"Referencing—citing sources—may be a good example of something that can be offloaded to AI, he said. . . ." Good example of why you always start with the problem you're trying to solve and don't go in search of problems for a predetermined solution. Citation managers already do this very well.

insidehighered.com

AI: Cheating matters but redrawing assessment "matters most"

Universities should prioritize ensuring that assessments are “assessing what we mean to assess” rather than letting conversations be dominated by discussions around cheating.

Dave Bloom @drumkitdave.bsky.social · 24/12/2024

Excellent list! Would also recommend The Thin Man (1934) for both Christmas adjacency and William Powell-Myrna Loy chemistry.

Dave Bloom @drumkitdave.bsky.social · 14/12/2024

Can you say more about 'you want to know what some objections to an idea might be?' I'd think that to judge whether an objection to an idea is credible (rather than just possible), you need a source to evaluate, meaning this info need, too, would benefit much more from search than generative AI.

000

3. Making searching and info lit best practices resonate with people isn't easy even in an academic environment where librarians can actually teach the stuff. People take cues from the tools, and if the tools encourage sloppy use, any counter programming is at a huge disadvantage. /end

010

2. No rhetoric around generative AI best practices like 'you just need to get good at prompting' will overcome the messaging and appearances of the apps that encourage search engine-like use.

100

Failing generative AI isn't noteworthy, but notice these things: 1. People have poor search and source evaluation skills. People seldom go past the first page of results and most would absolutely take that initial, unsourced output as credible.

100

And, finally, I point out that the source it used was not one of the recommended sources, and it tries again, does poorly.

Input: "Doesn't look like that data came from IDC, Garner, or Statista." Output: "You're right."+more nonsense and more bad sources that are not very illuminating

100

And I ask it to look up real data in the sources mentioned and create a pie chart from it. Here's the source it provides: www.oberlo.com/statistics/u... And since that source isn't much of a source, here's the source for the source: gs.statcounter.com/vendor-marke...

Input: "Yes, do that. Then make a pie chart of the data" Output: "I found the latest data on smartphone market share in the United States for 2024. Here is the updated pie chart based on the data" plus pie chart with Apple at 57.39%, Samsung at 23.27%, and Others at 19.34%.

100

The output - a chart with no sourcing!

Prompt as described in previous, output = "here's a pie chart showing the market share of smartphones in the United States, with a pie chart showing the following: Apple - 55%, Samsung - 25%, Google - 10%, OnePlus - 5%, Others - 5%. If you need any further details or adjustments, feel free to let me know!"

100

Here's the list of suggested prompts currently included in the UW's institutional instance of Copilot. I went with the pie chart suggestion.

List of Copilot suggested prompts, including "Graph the data quickly: Create a pie chart showing the market share of smartphones in the United States"

100