Generative Artificial Intelligence in the Libraries
by Juan Denzer, Engineering and Computer Science Librarian
Generative Artificial Intelligence (AI) has gained attention in academia beyond the ethical use of AI among students. Researchers and developers have been working to create and enhance their AI tools specifically for academic researchers to provide a better and easier user experience. Syracuse University Libraries is committed to using, testing and providing research guidance on the use of these AI tools in the research process.
Generative AI assistants, like ChatGPT, Microsoft Copilot and Google Gemini, are for public use but are also being used to develop advanced AI tools for academic research. Tools such as Keenious and Perplexity.ai are focused on providing research experiences similar to how people in academics search for resources. They are designed to help faculty and students use AI to gather information for literature reviews relevant to their topic and tools that provide summaries, citation count, citations and resource management.
Such tools for a better research experience are not new for library services. Databases such as Scopus, Dimensions, Web of Science and Inspec Analytics Plus already provide algorithm-based searches that include analytics for citation counts, relevant resources, management and data visualization. The use of Generative AI continues to evolve to provide a much better researcher experience.
Providers of resources such as Web of Science, Scopus, Dimensions and ProQuest One Literature are developing new Generative AI tools that will be incorporated into their databases. These tools will have a huge advantage over AI tools such as Keenoius and Perplexity.ai, because they can search data that is not limited to Open Access (OA). The current generative AI tools are crippled by being limited to access abstracts and full text only from OA publications, with access to abstracts from non-public publications also limited.
Here are some of the most popular information providers working on new generative AI tools for research and their features:
Web of Science (WoS) Research Assistant, trial access to this resource is provided through December 31, 2024.
- Ask questions based on a topic, journal or researcher.
- Multiple language support for questions and responses. Users not familiar with English will be able to ask their questions and receive the returned summary in their native language.
- Present the most relevant research to the user's question, with an overall commentary of papers distilled from article abstracts.
- Visualizations of documents on a topic over time.
- Unique data visualizations that show how output on a particular topic has varied over time.
- Context-specific visualization.
- Citation connections visualization.
- Topic map visualization.
- User interface design that is like ChatGPT and Google Gemini, making it easier to navigate.
- Throughout the trial period, contact Juan Denzer with any questions on using the tool or feedback for Web of Science on the development of this resource.
Scopus AI, trial access to this resource is provided through December 3, 2024.
- Draws from metadata and abstracts of Scopus documents published since 2013.
- Reliable and digestible research summarization from article abstracts.
- Provides the references used to build those summaries.
- Identify influential voices in a field; the Foundational papers option lists the most high-impact Scopus papers on any topic.
- Data Visualization; Scopus AI uses keywords from research abstracts to generate a concept map.
- Topic experts feature draws on the 19.6+ million author profiles in Scopus to find the top researchers linked to your query and generate a summary of their work and contributions.
Dimensions AI Assistant from Digital Science
- Provide contextualized synopses using what is known as extractive and abstractive summarization via Dimensions AI Summarization, included in the current subscription to Dimensions.
- Scans 33 million articles from the Dimensions database to retrieve and semantically rank the top results.
- Abstracts of the top four results are processed by Open AI’s GPT Models API to generate abstractive summaries.
- Dimensions General Science-BERT model is used to provide extractive answers for the top 10 publications.
ProQuest One Literature Research Assistant
- Humanities focused AI research tool.
- Provides a chatbot experience like other Generative AI Assistants, such as ChatGPT and Gemini.
- Draws from metadata and abstracts from ProQuest documents.
- User interface design like ChatGPT and Google Gemini, making it easier to navigate.
Syracuse University Libraries is excited to be a leader in providing research, education and feedback for these new advances in generative AI. And we are committed to helping our campus community in their research. We encourage our faculty, students and staff to take advantage of these opportunities to provide feedback for developers of these AI tools.
Please note, these resources, and all Libraries’ provided resources are for use in alignment with the Licensed Web Resources Policy. Uploading or otherwise incorporating licensed E-resource content into AI systems (University authorized as well as other) may be restricted or governed by individual license agreements that vary by vendor and publisher. Please consult with your subject librarian before using any Libraries’ licensed content for this purpose.