 Your new post is loading...
 Your new post is loading...
|
Scooped by
heather dawson
July 23, 2025 6:52 AM
|
Objective: As data science becomes more integrated into research and teaching, libraries are well-positioned to support this work. This study examines how a digital scholarship team at Binghamton University enhanced engagement with data science by assessing faculty, staff, and graduate student needs. Through focus group interviews, the study identifies key support gaps and outlines strategic initiatives to strengthen interdisciplinary data science programming within the library. Methods: A qualitative approach was used, involving 26 focus group interviews with faculty, staff, and graduate students across STEM and related fields. Participants discussed their data science work, tools, training, and perceived resource gaps. Qualitative coding analysis identified key areas for library support. Results: The study revealed three primary areas for library expansion in data science: (1) fostering interdisciplinary collaboration through outreach, (2) developing structured data science programming aligned with campus needs, and (3) establishing physical and digital infrastructure for data-intensive research. In response, the Digital Scholarship team implemented a three-semester data science programming plan, enhanced research community engagement, and contributed to a dedicated data science space in the upcoming Digital Scholarship Center. Conclusions: Findings support the library’s role as a vital hub for data science. By aligning digital scholarship services with campus needs, the library can bridge gaps in data literacy, tool accessibility, and collaborative opportunities. While initial implementations show promising engagement, ongoing assessment will be necessary to refine services, particularly for undergraduates and emerging technologies. This study provides a model for other libraries to expand data science programming effectively.
|
Scooped by
heather dawson
July 20, 2025 4:13 PM
|
Developments in Artificial Intelligence markets: New indicators based on model characteristics, prices and providers
|
Scooped by
heather dawson
July 20, 2025 4:11 PM
|
|
Scooped by
heather dawson
July 20, 2025 4:07 PM
|
Large Language Models (LLMs) have shown promise in accelerating the scientific research pipeline. A key capability for this process is the ability to generate novel research ideas, and prior studies have found settings in which LLM-generated research ideas were judged as more novel than human-expert ideas. However, a good idea should not simply appear to be novel, it should also result in better research after being executed. To test whether AI-generated ideas lead to better research outcomes, we conduct an execution study by recruiting 43 expert researchers to execute randomly-assigned ideas, either written by experts or generated by an LLM. Each expert spent over 100 hours implementing the idea and wrote a 4-page short paper to document the experiments. All the executed projects are then reviewed blindly by expert NLP researchers. Comparing the review scores of the same ideas before and after execution, the scores of the LLM-generated ideas decrease significantly more than expert-written ideas on all evaluation metrics (novelty, excitement, effectiveness, and overall; p < 0.05), closing the gap between LLM and human ideas observed at the ideation stage. When comparing the aggregated review scores from the execution study, we even observe that for many metrics there is a flip in rankings where human ideas score higher than LLM ideas. This ideation-execution gap highlights the limitations of current LLMs in generating truly effective research ideas and the challenge of evaluating research ideas in the absence of execution outcomes.
|
Scooped by
heather dawson
July 20, 2025 3:57 PM
|
|
Scooped by
heather dawson
July 16, 2025 3:27 AM
|
View examples of cases we receive and how we resolve them. These include our public interest cases and case summaries.Office of the Independent Adjudicator for Higher Education
|
Scooped by
heather dawson
July 11, 2025 4:23 AM
|
Data rescue project new website DRP Portal— over 1,000 datasets rescued from American government depts and agencies by volunteers. Lok by dept and dataset and theme
|
Scooped by
heather dawson
July 2, 2025 3:30 AM
|
Evaluating and prioritising artificial intelligence projects A guide for better decision making and investment outcomes 1 Jul 2025 Publisher
|
Scooped by
heather dawson
June 27, 2025 4:05 AM
|
How NHS communications professionals are adopting AI tools and what is needed to enable safe, ethical and confident use.
|
Scooped by
heather dawson
June 19, 2025 4:13 PM
|
This document brings together the main advances driven by AI, a powerful tool that offers practical solutions for university faculty, students, researchers, and academic institutions
|
Scooped by
heather dawson
June 12, 2025 11:24 AM
|
This paper evaluates the performance of six open-weight LLMs (llama3-8b, llama3.1-8b, gemma2-9b, mixtral-8x7b, llama3-70b, llama3.1-70b) in recommending experts in physics across five tasks: top-k experts by field, influential scientists by discipline, epoch, seniority, and scholar counterparts. The evaluation examines consistency, factuality, and biases related to gender, ethnicity, academic popularity, and scholar similarity. Using ground-truth data from the American Physical Society and OpenAlex, we establish scholarly benchmarks by comparing model outputs to real-world academic records. Our analysis reveals inconsistencies and biases across all models. mixtral-8x7b produces the most stable outputs, while llama3.1-70b shows the highest variability. Many models exhibit duplication, and some, particularly gemma2-9b and llama3.1-8b, struggle with formatting errors. LLMs generally recommend real scientists, but accuracy drops in field-, epoch-, and seniority-specific queries, consistently favoring senior scholars. Representation biases persist, replicating gender imbalances (reflecting male predominance), under-representing Asian scientists, and over-representing White scholars. Despite some diversity in institutional and collaboration networks, models favor highly cited and productive scholars, reinforcing the rich-getricher effect while offering limited geographical representation. These findings highlight the need to improve LLMs for more reliable and equitable scholarly recommendations.
|
Scooped by
heather dawson
June 12, 2025 11:23 AM
|
In the aftermath of the “AI boom,” this report examines how the push to integrate AI products everywhere grants AI companies - and the tech oligarchs that run them - power that goes far beyond their deep pockets.
|
Scooped by
heather dawson
June 12, 2025 9:11 AM
|
This ILO Working Paper refines the global measurement of occupational exposure to generative AI by combining task-level data, expert input, and AI model predictions. It offers an improved methodological framework to assess how GenAI may impact jobs across countries and sectors.
|
|
Scooped by
heather dawson
July 20, 2025 4:19 PM
|
The purpose of these meetings was to identify key issues, brainstorm achievable solutions, and then prioritize a plan of work for NISO to improve efficiency working with AI systems across the network of publishers. In each workshop, participants generated and refined dozens of ideas for possible standardization projects and then chose the three project ideas with the greatest value to the broader ecosystem. The discussions focused on interoperability, licensing, attribution, data provenance, transparency, accessibility, and evolving usage models.
The resulting report, Advancing Artificial Intelligence Interoperability in STM Content through Standards, includes a detailed narrative summary of each workshop and a one-page synthesis of outcomes from each session. It concludes with a consolidated list of nine strategic standardization opportunities. These areas represent high-impact intersections between AI, content infrastructure, and scholarly integrity, and will be evaluated by NISO for potential leadership, collaboration, or support.
|
Scooped by
heather dawson
July 20, 2025 4:12 PM
|
|
Scooped by
heather dawson
July 20, 2025 4:08 PM
|
Web-enabled LLMs frequently answer queries without crediting the web pages they consume, creating an “attribution gap” – the difference between relevant URLs read and those actually cited. Drawing on approximately 14,000 real-world LMArena conversation logs with search-enabled LLM systems, we document three exploitation patterns: 1) No Search: 34% of Google Gemini and 24% of OpenAI GPT-4o responses are generated without explicitly fetching any online content; 2) No citation: Gemini provides no clickable citation source in 92% of answers; 3) High-volume, low-credit: Perplexity’s Sonar visits approximately 10 relevant pages per query but cites only three to four. A negative binomial hurdle model shows that the average query answered by Gemini or Sonar leaves about 3 relevant websites uncited, whereas GPT-4o’s tiny uncited gap is best explained by its selective log disclosures rather than by better attribution. Citation efficiency – extra citations provided per additional relevant web page visited – varies widely across models, from 0.19 to 0.45 on identical queries, underscoring that retrieval design, not technical limits, shapes ecosystem impact. We recommend a transparent LLM search architecture based on standardized telemetry and full disclosure of search traces and citation logs.
|
Scooped by
heather dawson
July 20, 2025 4:06 PM
|
Latest News Kathryn Dally joins UKRIO as Director of Programmes UKRIO publishes revised Code of Practice for Research UKRIO publishes new guidance to support responsible AI use in research UKRIO publishes summary of authorship workshops UKRIO publishes summary of 2025 subscriber symposium New...
|
Scooped by
heather dawson
July 18, 2025 4:35 AM
|
The Charity Digital Skills Report is the annual barometer of digital skills, attitudes and support needs across the sector.
|
Scooped by
heather dawson
July 16, 2025 3:11 AM
|
This report is the output of a series of four workshops in Summer and Autumn 2024. The workshops brought people together from across disciplines and practices to discuss how artificial intelligence (AI) can work in society for the common good and, in pursuit of that objective, to find gaps in the research agenda informed by policy needs. The report summarises discussions about applying AI for the public good across four thematic areas – administrative justice, place, public health and market failure – and presents the observations, ideas and research questions generated in the workshops. The report concludes by identifying how AI may contribute to, or hinder, improvement of services and outcomes for citizens. It identifies a series of challenges and research questions the arose in each of the four workshops, including: - supporting decisions about applying AI
- collaboration and collective learning for co-design
- appropriate uses of automation
- understanding relevant current use of AI
- data management and trust
- understanding legitimacy.
|
Scooped by
heather dawson
July 8, 2025 4:14 AM
|
Public attitudes to public sector AI
|
Scooped by
heather dawson
June 27, 2025 3:22 PM
|
https://lustre-network.net/about/ this project has involved academics, AI specialists, records managers ,museum and archives roles to explore the potentials of AI in their sector in the future the website provides free access to their publications and workshop outputs including: And a special recent journal article collection according to the project ‘This collection demonstrates the transformative role of AI in advancing archival practices across government and GLAM sectors. The articles showcase how AI can unlock digital data while considering issues such as bias, privacy, and trust.”.
|
Scooped by
heather dawson
June 19, 2025 4:15 PM
|
By leveraging collective intelligence and scenario planning, strategic foresight exercises can help ensure that AI technologies are developed and deployed responsibly, thereby increasing the likelihood of their positive contribution to sustainable and inclusive growth. Such forward-thinking methodologies are critical to mitigating risks and harnessing AI’s transformative power in advancing the SDGs.
|
Scooped by
heather dawson
June 17, 2025 3:15 AM
|
Digital literacy and inclusion critical to skilling individuals and businesses for success in the age of AI.To ensure individuals, businesses, and communities can thrive in the age of AI, government policy must place inclusion, digital skills, and adaptive regulation at the heart of a thriving, future-ready society.
|
Scooped by
heather dawson
June 12, 2025 11:23 AM
|
To develop a robust policy for generative artificial intelligence use in higher education, institutional leaders must first create "a room" where dive
|
Scooped by
heather dawson
June 12, 2025 9:20 AM
|
An unprecedented look at the state of AI’s energy and resource usage, where it is now, where it is headed in the years to come, and why we have to get it right.
|