Kerri Lemoie, PhD

This. Thank you for guiding all libraries on this topic. Imagine having a disability that doesn’t allow you to access the information you’re seeking. Please don’t rush the process. Allow yourself to be informed and make sure you’re making the right decision for the communities you represent. Not all tools are made the same, and some offer a different user experience. It is essential to prioritize accessibility options for individuals with disabilities on your website. This should not be viewed as merely a compliance issue but rather as critical as your collection development, weeding practices, and programming. Ensuring accessibility impacts how users conduct searches on your site. Happy Monday. #accessibilitymatters

6

1 Comment

Join the conversation on AI in libraries! Submit your proposal for OhioNet’s "AI Alchemy: Libraries Growing Smarter" summit by 9/13: https://loom.ly/3vEWueE. 💡 We're seeking insights on how AI is transforming library services, enhancing user experiences, and driving growth.

2

This literature review explores the history of libraries, current laws, and studies. It examines digital accessibility and physical accessibility issues impacting libraries. A major issue is library webpages and PDFs having accessibility errors. This article overlooks third-party library databases. I hope that more academic journals and databases offer their published works as accessible HTML pages. This can make library resources accessible to all. #accessibility https://lnkd.in/gBfDrpbi

4

How much should #k12 and #highered consider the #enterprise editions of the #frontend #genAI and #LLMs? Some #education institutions secure campus-wide #access for their #students, #faculty, #administrators, and #staff to #GenerativeAI. How is the rest of the educational world providing access to their employees? How is the educational world copying with the extremely rapid changes in the offerings of frontend platforms, considering the latest version of #Anthropic's #Claude? EP 351: 3 Reasons Businesses Shouldn’t Use Anthropic's Claude (Yet) Everyday AI Podcast – An AI and ChatGPT #Podcast https://lnkd.in/dN_EDdpd

2

International collaborations and library funding are critical for transitioning specialized book collections into #OpenAccess as a public good, while significantly augmenting their visibility and usage https://ow.ly/uVY150TXm6q

1

International collaborations and library funding are critical for transitioning specialized book collections into #OpenAccess as a public good, while significantly augmenting their visibility and usage https://ow.ly/CMm350TXm0A

5

Big Ten Academic Alliance (BTAA) libraries are now live using the new OCLC Resource Sharing for Groups, the state-of-the-art solution that reimagines resource sharing for library groups and consortia. OCLC Resource Sharing for Groups is a modern consortial borrowing solution that expedites delivery of digital and print items to library users. Participating libraries have all the tools needed to easily lend and borrow resources from group members. "We are pleased with the successful migration to this new platform for UBorrow—not only for the parity and increased robustness it brings our communities of users today, but for the expanded options and possibilities we envision as we continue our pathway to uniting our separate collections into one, fully shared and fully networked," said Maurice York, BTAA Director of Library Initiatives. "We consider this new platform a fundamental part of our infrastructure renewal, a key component in preparing us to expand this fall to become a coast-to-coast Alliance." ⤵️ Click below to read more. #OCLCnews --- #OCLC #ResourceSharing #InterlibraryLoan #PublicLibrary #PublicLibraries #AcademicLibrary #AcademicLibraries #Library #Libraries #Librarian #Librarians #LibraryLife #LibraryLove

46

1 Comment

This webpage profiles risk statements related to accessibility. Some of the risk statements relate to an institution failing to assign a department or person to coordinate institution-wide accessibility, failing to provide accessible instructional materials in a timely manner, and failing to provide an accessible website. Specific cases are referenced below each risk statement, which can provide context. #accessibility https://lnkd.in/g5fBNGGC

3

This is an interesting dissertation that explores digital accessibility by analyzing higher education institutions in Colorado. The study finds the most common digital accessibility barriers on the websites of public and private institutions. Private and vocational schools had more accessibility errors. The overall error rate (2.5%) was calculated by comparing a small fraction of inaccessible elements to all the web elements on a webpage. #accessibility https://lnkd.in/g2_i6mJz

3

Here's the first of 3 short articles I've written for ACRL's LibTech Insights: https://lnkd.in/gtAKWSPV Next week will focus on ethical questions and the following week, practical questions.

17

1 Comment

New guidance docs released today about establishing #ai policy in #k12 from TeachAI https://lnkd.in/eaBKbiM2

4

If you are doing design science reserch on artificial intelligence topics then have a look at this. Of course, this is interesting to anyone who wants to understand how to approach design-oriented research on AI.

21

1 Comment

How are libraries approaching #AI? What are their key concerns and challenges? What are the implications for the future role of librarians and library patrons? Learn what over 1,500 librarians from academic, national and public libraries surveyed by Clarivate had to say in the 2024 Pulse of the Library report. Read our blog here: https://lnkd.in/etctrjCu

2

I consider myself part of the open education community, going back all the way to the launch of OpenCourseWare. What “open” means in the context of AI is not clear. The fundamental rights of open source software don’t translate easily, at least not to the most powerful foundation models, because the cost of running them makes it infeasible for most. And inspecting what goes on “inside the black box” is much harder than reading software code. However, for me “open” has always been most interesting as a means to an end: increasing opportunity for meaningful participation. And I am curious how that spirit could translate to AI. I very much look forward to the collection of articles that this project will generate. And if you have a perspective or a recommendation you want to share, please submit a proposal!

60

4 Comments

International collaborations and library funding are critical for transitioning specialized book collections into #OpenAccess as a public good, while significantly augmenting their visibility and usage https://ow.ly/jkQn50TXm3g

1

I got a question from a member librarian this morning asking about AI detection sites. This was my answer: The rate of false positives, combined with false negatives, means the actual detection success is about like flipping a coin. I have seen an increase in detection reliability for simple prompts and informational writing (write an essay about this topic) but outside of that, ChatGPT’s 4o is making detection more difficult. OpenAI, the company behind ChatGPT, has had internal discussions about watermarking text. They are most concerned about the equity issues around this as it would disproportionally impact certain groups like English language learners and those with learning challenges [https://lnkd.in/gX2anaE5]. Another issue is that most of the detection tools also happen to be selling other writing tools at the same time. For example, Quillbot has a detector, but also has a summarizer tool that uses AI. Another tool, Scribbr, also has AI writing tools. Seems like a bit of a conflict of interest to me when the so called solutions are also selling the problem that they claim to be solving! If we are going to be ethically sketchy, let’s at least be clear about it. I love https://undetectable.ai/. It runs a meta-checker that checks text vs a range of detection sites to give you a better overview. It also shows you why this is never going to be a real solution because they are in an arms race with the detectors. Undetectable sells an AI Humanizer that re-writes AI content to remove the markers that may reveal it to be AI writing. They also offer a human typer tool that slowly pastes text into a Google Doc to bypass checks on large paste dumps in the history of the document. Which brings us back to the equity issue identified by OpenAI. Yes, they could watermark text, but those with the means to pay $60 a year to cheat will still do so. To paraphrase War Games, the only way to win is not to play the game. A better approach would be to rethink how we assign and assess work. Math teachers had to modify their instructional practices over the last decade to address Photomath. They assign fewer problems, focus more on the understanding of the process than the final answer, and work to prepare students for assessments that score the work as well as the answer. If the goal in ELA is to assess writing skills, this can be done with a 15 minute quick-write in class. That will give the teacher enough of a writing sample to assess technical skills. If the goal is to evaluate the creation and internalization of an argument based on deep understanding and synthesis of a topic (a research paper) then assess it via a performance more similar to a dissertation defense. This is about making sure that the objective of the assessment is aligned with the tool. Using a 10 page paper to assess technical writing skills is a poor alignment.

17

1 Comment

Just read this interesting article from Inside Higher Ed "Is AI Finally a Way to Reduce Higher Ed Costs?" Written by José Antonio Bowen & C. Edward Watson Very interesting perspectives on the impact of AI in higher ed. There is a lot of functions AI can potentially serve in higher ed such as course design, assessment, student support, grant writing and more. While I don't think AI will be able to replace the human factor and the need for human oversight. I wonder will universities lower staff count or keep similar faculty numbers with changed responsibilities. There is a lot of talk in general when it comes to AI that it will replace existing job roles. But I think it will create new job roles also. Will it create more jobs than it destroys? What do you think? Link to the article 👇 https://lnkd.in/e43UZjFN #higheredmarketing #highereducationmarketing #highered

10

4 Comments