Scopus AI is a search tool developed by Elsevier that uses generative AI to provide suitable responses to technical search terms, even answering in whole sentences. Depending on the input, Scopus AI is suitable for finding sources for scientific papers, gaining an initial overview of a research topic, or deepening your understanding of it, articulating an idea or hypothesis, and explaining a (complex) subject to an audience.
Functions
When you enter a technical term or full sentence (in English or another language), Scopus AI pairs this input with the most relevant titles and abstracts from the Scopus database that best match the query. Scopus AI then presents the following results:
- Summary: a short response to your input, including the Scopus sources used.
- References: a list of the Scopus sources used in the summary.
- Concept map: a downloadable concept map is created for each search query using keywords from abstracts. This visual mapping of search results can identify connections between research topics. Watch this video about concept maps.
- Expanded summary: a more detailed version of the summary.
- Foundational papers: a list of high-impact papers compiled with reference not only to their number of citations but also the relevance of the content to your question.
- Topic experts: a list of experts in the respective research field selected from the Scopus database, with a summary of their relevant expertise, as well as their Scopus metrics (h-index, number of documents matching the input and their total citation score).
- Deeper questions: suggestions for related search queries (prompts). It can include similar technical terms and thus expand the user's technical vocabulary.
How Scopus AI searches
One advantage of Scopus AI is that it lists the sources used along with the output. The link between sentences and sources is based on relevance determined by a so-called vector search. In contrast to conventional Boolean searches, which match exact terms, a vector search takes into account the contextual meaning of the content. This allows Scopus AI to identify sources that are semantically related. The result is that contextually relevant sources can be included in the output, even if none of the terms used in the input appears directly in their abstract.
The concept map is generated using Large Language Models (LLM).
Scopus AI uses a technology called RAG Fusion, developed by Elsevier, to create expanded summaries. RAG Fusion summarises the question entered and performs a compact systematic literature review.
Elsevier uses the citation graphs it has created and refined over the years to list foundational papers that match the search queries.
The topic experts list is created using Elsevier's Researcher Discovery tool.
Data corpus
When creating suitable responses, Scopus AI accesses not only titles but also the abstracts of documents published from 2003 onwards using the multidisciplinary Scopus database. Scopus AI is also updated as soon as the Scopus database is updated. A positive side effect is that this limits the inclusion of content from retracted articles in the Scopus AI responses.
Information on the Scopus database in DBIS
Access
You can access Scopus AI from the Scopus search home page by clicking the [Scopus AI] tab. Please note that use is voluntary, and the University Library is not responsible for information processing within the AI tool. The Scopus AI module will be licensed in 2025 by the University Library (as part of the Austrian Academic Library Consortium (Kooperation E-Medien Österreich)) for the University of Graz in an extended test phase. Please pay attention to both the Elsevier and University Library Graz terms of use.
The database is accessible from anywhere on campus, and university members can also access it remotely via VPN. Please note: enter https://univpn.uni-graz.at/ub in the VPN address field.
Have you discovered our video tutorials on unikat, literature research and literature management? We also offer training courses during the semester. You can find more information under Support & Courses.
If you have any questions about this database, please contact ub.zeitschriften(at)uni-graz.at.