I'm also trying to integrate generative AI into an undergraduate metadata course. We discuss AI-generated subject metadata for text and images and how to evaluate it. But we start with core metadata skills.
How can generative AI be implemented in the metadata curriculum?
doi.org/10.47989/ir3...
Why read? A presentation of early experiences in incorporating AI-related elements into teaching and learning.
What’s next? More articles like this are needed.
#ShortReview #AI #Curricula #Metadata #Teaching
24 kwietnia o godz. 18 w Centrum Nauki Kopernik odbędzie się spotkanie z Karen Hao, autorką książki "Imperium sztucznej inteligencji" (wstęp wolny). Wydarzenie będzie także transmitowane online:
www.youtube.com/watch?v=SYNE...
#AI #CentrumNaukiKopernik #KarenHao
Thank you! I'll look into her thesis. I also noticed that earlier volumes were indexed with the descriptor "Rabbit."
A Japanese hare seems more appropriate. This is a small example of the problems encountered in subject indexing of fictional works.
I'm reading Stan Sakai's Usagi Yojimbo series. The subject description for the volume Crossroads in the National Library of Poland catalog uses the descriptor "European rabbit." But is that really the best term for Usagi Yojimbo?
#Comics #UsagiYojimbo #Indexing #LIS #KO
From a knowledge organization perspective, the framework uses GraphRAG based on knowledge graphs to answer thematic queries; however, the paper also points to the limitations of this solution.
#AI #GraphRAG #KnowledgeGraphs #KO
In my view, one of the most significant directions for future research is the hybrid model mentioned by the authors. Supporting human interaction is an important mission of libraries.
#AI #Libraries #LIS #Users
How can AI support reference services in a special library?
doi.org/10.1108/EL-0...
Why read? A practical AI solution for reference services & comprehensive evaluation.
What’s next? Authors: exploration of hybrid models (AI & humans).
#ShortReview #AI #SpecialLibraries #LIS #ReferenceServices
New Research: Have LLM-Associated Terms Increased in Article Full Texts in All Fields? (via #arXiv) arxiv.org/abs/2604.07565 #LLMs #GenAI #AI #scholcomm #journals
For me, as a LIS/KO researcher, the documentary is one more piece of evidence that we should focus less on the technological side of AI, and more on its ethical and social aspects and its impact on humans as information agents and on information-related organizations.
#LIS #KO #AI #Research
I've finally watched the documentary "The Thinking Game", a portrait of Demis Hassabis and the DeepMind laboratory. The film is really worth watching as a point of departure for thinking about the future of AI.
www.youtube.com/watch?v=d95J...
#AI #DeepMind #DemisHassabis #Documentary
The authors rightly emphasize that their study shows to what extent chatbots may be used in smaller cultural institutions with limited resources (p. 418). This is an important contribution, with implications for many smaller organizations.
What is the quality of ChatGPT-created metadata for objects from a community archive?
doi.org/10.47989/ir3...
Why read? Comprehensive evaluation framework & informal prompting test.
What’s next? Authors: more developed prompts are needed.
#ShortReview #AI #ChatGPT #Metadata #CommunityArchives
At a media forum in Zhengzhou last weekend, foreign influencers praised China's "authenticity," while AI was celebrated as the ultimate tool for producing "positive energy." No hand-wringing about trust or ethics here. Just propaganda, supercharged.
linguasinica.substack.com/p/ai-for-hum...
I’ve published a monthly annotated bibliography on knowledge organization, metadata, and AI. It includes 12 articles and preprints, each with a brief commentary.
doi.org/10.6084/m9.f...
#AI #Metadata #KO #Bibliography #AIKOBib
The NKOS Workshop will be held at DCMI2026 in Seoul. The call for submissions is open until April 30, 2026.
nkos.dublincore.org/NKOS2026CFP....
#AI #KO #DCMI #Conference #Korea
In my view, such papers show a real-world scenario. At this stage of AI implementation in libraries, librarians can use chatbots through a GUI.
#AI #Libraries
One may claim it is not a fully controlled experiment, since the model itself is not fully controlled. Yet the authors highlight that it is an exploratory study.
Can ChatGPT generate MARC 21 records for Arabic resources?
doi.org/10.1080/0163...
Why read? It presents prompt development & evaluates AI output for Arabic resources.
What’s next? Larger-scale research & improvements in evaluation strategy are needed
#ShortReview #AI #ChatGPT #Cataloging #Arabic
Company News: New Clarivate Study Examines the Impact of Academic #AI on Core Library #Workflows
Report: clarivate.com/academia-gov...
Release/Highlights: clarivate.com/academia-gov...
#metadata #libraries @clarivateag.bsky.social
Here’s Anthropic’s official guide to Claude Skills: resources.anthropic.com/hubfs/The-Co...
#Claude #Skills
Jamie Keet (Teacher’s Tech) has published a video on Claude Skills. It’s an interesting example of how AI customization can support task-specific user needs.
youtu.be/mS5ojqQ7zzw?...
#AI #Claude #Skills
Jakie miejsce zajmuje AI w nowym planie 5-letnim ChRL?
Eksperci Zespołu Chińskiego OSW omawiają wybrane założenia dokumentu i podkreślają, że sztuczna inteligencja ma wspierać gospodarkę oraz rozwiązywanie problemów społecznych.
youtu.be/ic6iot_Zt2s?...
#AI #Chiny #Plan5Letni
DeepSeek R1 models struggle with theme extraction. The paper suggests that more abstract interpretation remains difficult for smaller LLMs. There is still a clear role for human expertise here.
#AI #LLM #DeepSeek #HumanExpertise
How do language models perform at extracting topics and themes from comments?
doi.org/10.23106/dcm...
Why read? A useful example of how IR metrics can be used to evaluate topic and theme extraction.
What’s next? Authors: LLM methods still need refinement.
#ShortReview #AI #LLM #Topics #Themes
The 2025 COS Impact Report is now live! From open infrastructure to research initiatives, policy engagement, and global partnerships, see what COS accomplished over the past year to advance transparent, rigorous, and accessible research.
🔍 Check out the report: www.cos.io/impact
Chińska bezpieka i ChatGPT
https://wp.me/p3fv0T-i7s #Japonia #Chiny #ChRL #dezinformacja #bezpieka #OpenAi #AI #Qwen #DeepSeek #ChatGPT #POLECANE
🔥 Generative AI is reshaping scientific research. But do these systems actually understand science? In our new paper, we peek inside LLMs to understand how they encode scientific quality, revealing several monosemantic features that predict research quality.
📄 arxiv.org/pdf/2602.19115
Second, the author shows that the addition of synthetic data considerably improves the values of metrics such as F1 score or recall@5. It can be leveraged in other studies and in practice.