JSTOR AI Research Tool:
JSTOR's AI research tool is currently in a beta phase to evaluate its usefulness and efficiency. When you search in JSTOR and click on a publication, you will now see the [AI Research Tool] button on the right, which you can set to [ON]. The AI research tool will then open next to the article, which, in addition to its ‘chat’ function, can:
show how a publication chosen from the search results list (article, chapter) relates to your search term (content relevance),
answer questions about the content of the article (key ideas, etc.),
search for similar articles in JSTOR when you click [Show related content],
suggest suitable search terms for further searches when you click [Related topics].
Please note:
The responses generated are based on content on the JSTOR platform.
To use this tool, you first need to register for a free JSTOR account.
EBSCO AI Insights:
To gather user feedback, EBSCO launched another beta test of its AI tool AI Insights in July 2025 (see the official EBSCO press release). Researchers using the EBSCOhost platform can test the AI Insights feature, which generates a list of "insights" from the full text of a selected publication. The ‘insights’ are a kind of summary of the article, which is intended to help younger researchers in particular to choose their reading and literature by using more accessible language, structure and explanations. To change the platform language to English, click on [MyEBSCO] at the top right, then "Meine Einstellungen". Select English (or one of the other 32 languages available!) from the first dropdown menu and then click [Speichern] at the bottom of the page.
How to access content compatible with the tool:
As the tool uses the full texts of publications, it only works with publications that are part of databases licensed by the University of Graz.
Click on [Publications authority] on the left-hand side of the EBSCOhost platform. Then click on a database with a green ‘PDF’ icon next to it. Finally, click on a [Year] and then on an [Issue] to access the articles in that issue. The [Create AI Insights] button will appear next to the article entries. Finally, click on a year, and then on an issue to access the articles in that issue. The [Generate AI Insights] button will appear next to the article entries.
Please note:
Insights are regenerated with each request and are not reproducible.
The tool cannot generate insights from publications that are too short.
Although EBSCO has tested the tool's accuracy with real test subjects, users should still verify the insights generated for themselves.
Reaxys AI Search:
Elsevier has also launched the beta phase of its Reaxys AI Search tool. University of Graz members can log in (login guide) to access the [Reaxys AI Search Beta] tab. Reaxys AI Search uses vector-based searches and acts as an agent that can understand input in natural language. This tool is primarily intended for researchers looking for chemistry literature. It is also designed to make finding literature from related sub-disciplines easier.
Please note:
Data base: the whole content of Reaxys
Reaxys AI Search Beta does not learn from, nor is it trained on, users' queries or search results.
This early version does not generate summaries or interactive responses. It only finds relevant literature-based results, supporting accurate and efficient discovery.
As of September 2025, generated results lists contain a maximum of 100 entries.
Each entry in the results list is assigned a confidence score between 0 and 1 (1 = high confidence that the document is relevant ; 0 = low confidence). The results list is sorted in descending order according to these confidence scores.
To sort entries from the results list by other parameters (e.g. relevance, publication year, document type), select the entries you want and click [View in Reaxys.com] to open the list on Reaxys.com, where you can sort it.
Access and terms of use:
JSTOR: www.jstor.org
EBSCO: research.ebsco.com/
Reaxys: www.reaxys.com
The three databases are accessible from anywhere on campus via unikat and DBIS; university members can also access them remotely via VPN. Please note: enter https://univpn.uni-graz.at/ub in the VPN client's address field.
Please pay attention to the JSTOR Terms and Conditions of Use, EBSCO Terms of Use, and Elsevier (Reaxys) Terms and conditions. The University Library accepts no liability for the processing of your data by the provider.
If you have any questions, please contact ub.zeitschriften(at)uni-graz.at.