Very wild behavio(u)r

a stuffed white cat is sitting on top of a box and asking why ? Alt: a stuffed white cat is sitting on top of a box and asking why ?

Additionally, when a passkey login still requires 2FA

Hey #Library folk check out this #LGBTQIA resources account for library, knowledge & information workers 📚📜

#MedLibs

Bringing libguides up to WCAG 2.1 AA standards has made me unfortunately aware of how often I include “click” in my written instructions 😬

Oh no! I hope you feel better soon 😕

a young boy in a hawaiian shirt is asking can we please get on the same page here ? ALT: a young boy in a hawaiian shirt is asking can we please get on the same page here ?

Increasingly disturbed to see large, important, many-authored international guidelines published without the involvement of any information specialists whatsoever. Guidelines involve literature searching, methodological knowledge, evidence synthesis, documentation, reproducibility & transparency.

I don’t think it’s saying “don’t use.” Instead, as with all effective literacy education, it’s about supporting users in making informed choices. I assume we both agree that contextualizing how AI generates content is necessary for any user to determine what utility the content has?

As political pressure mounts, medical school accreditor drops requirement to teach about health equity The leading medical school accreditation body in the U.S. has removed language from its standards requiring schools to teach about health inequities

Not being required to learn about very real health disparities will lead to less informed doctors (as a whole) and make it more difficult to address these inequities at a larger scale. #medlibs #medsky

In essence, all AI generated content are “hallucinations” - it’s just that some (or most) happen to reflect reality and some do not.

The primary challenge is that information literacy is aimed at understanding, discovering, and utilizing information. Generative AI creates probability patterns instead of information in its output. Most info lit framework adaptions rarely account for this distinction, and so are misaligned

Mid-Continent Public Library staff are unionizing. They could make Missouri history Librarians and staff in the Mid-Continent Public Library system have announced their intent to unionize. They want better wages and workplace protections. With more than 700 workers, it could be the l...

“If they win their union election, the Mid-Continent workers will be the second, and largest, library union in the state. The Daniel Boone Public Library, which primarily serves Boone and Callaway counties, was the first in Missouri to unionize in 2022.” www.kcur.org/news/2026-03...

GPT Tokenizer Playground – Visualize Tokens for OpenAI Models Interactive tokenizer playground for OpenAI models. Count tokens, estimate pricing, and learn how tokenization shapes prompts.

I’ve used sites like this to demonstrate to users how prompts are (in theory) assembled from their tokens without going into knowledge domains I can’t speak to gpt-tokenizer.dev

#medlibs

You might look into KeePass if you’re concern about local storage / access as an issue. Not the prettiest app but it’s usable. keepass.info

You are invited to participate in a survey on health sciences librarians’ use of generative AI (genAI) tools in their knowledge synthesis (KS) practice, including systematic and scoping reviews. Info & survey link: stfx.qualtrics.com/jfe/form/SV_...

#medlibs #canmedlibs #ukmedlibs #HealthLibAust

Indeed. Fëanor, by any other name, would still be an ass.

AI in UpToDate: New Generative Solutions for Medical Professionals As Artificial Intelligence (AI) expands into healthcare, building on top of trusted, expert, evidence-based content in UpToDate clinical decision support is crucial for safe, effective patient care.

#medlibs - have any been testing or developing further guidance on the usage of UpToDate’s Expert AI feature? www.wolterskluwer.com/en/solutions...

Controversial but - students *perceive* they have been failed when what has happened is they have failed to understand information retrieval is a skill requiring effort. I don't believe a zero friction experience is realistic or desirable in learning, actually.

Hallucinated citations produced by generative artificial intelligence may constitute research misconduct when citations function as data in scholarly papers In this article, we discuss the growing problem of hallucinated citations produced by Generative Artificial Intelligence (GenAI) in scholarly research and writing. We argue that GenAI hallucinated ...

Related: www.tandfonline.com/doi/full/10....

Others working in #evidencesynthesis and #systematicreviews may be interested 🙂

#medlibs - Version 3 of the Responsible use of AI in Evidence Synthesis (RAISE) is out! osf.io/fwaud/files/...

NIH Research Grants Drove A $94 Billion National Economic Impact In 2025 NIH research funding generated $94.15 billion in new economic activity across the nation in FY 2025, with a return on investment of $2.57 for every $1 invested.

"Research funding from the National Institutes of Health generated $94.15 billion in new economic activity across the nation in 2025, with a return on investment of $2.57 for every $1 invested."
www.forbes.com/sites/michae...

In many ways, it’s not even the original use of AI that shocks me - it’s that these ‘authors’ (in the loosest possible sense) don’t even bother to check what the AI has produced, despite how much we know about hallucinations. That’s a whole new level of pathological laziness. 📚

MAXTONE-GRAHAM v. BURTCHAELL, No. 92, Docket 86-7349. | 2d Cir., Judgment, Law, casemine.com Get free access to the complete judgment in MAXTONE-GRAHAM v. BURTCHAELL on CaseMine.

ex. 2 www.casemine.com/judgement/us...

this is not legal advice and without knowing particulars, there is precedent for fair use protection for criticism, but it depends on the extent and nature of the usage. ex. www.copyright.gov/fair-use/sum...

I recall an email hellthread from some years ago where someone mistakenly sent a Reply All to a mailing list message asking to be unsubscribed from replies, and subsequently many on the thread responded in kind not realizing they were perpetuating the thread. 🫠

Screenshot of new EBSCO interface change

“We’re moving the primary navigation from a left-hand panel to the top of the page, which creates a more consistent experience across screens and frees up space for content and tools.” assets.ebsco.com/m/635a707af2...