On this page:
JSTOR’s AI research tool empowers you to deepen and expand your research with JSTOR’s trusted corpus. Developed in collaboration with our community, this tool leverages AI and other advanced technologies to help you elevate your research, teaching, and learning.
If the research tool has been enabled for your institution, this beta feature will appear on the item page for journal articles, book chapters, and research reports, and as a search option in addition to JSTOR’s standard keyword search.
The tool is in active development so new features and improvements are frequently added. For more information on how the tool is powered and JSTOR’s approach to generative AI, check out our FAQs.
Evaluating and exploring content
Once you select an eligible text item from your search results, the research tool will start by identifying how your search query relates to the content by asking: “How is [your search query] related to this text?”
From here, you can continue evaluating the content through the research tool window or from the text itself.
Using prompts
You can learn more about the text as a whole by choosing one of the prompts provided by the research tool or by supplying your own. Read more about these options in the following chart.
| Prompt | Description |
| What is this about? / Show abstract |
Suggests the text’s key points and arguments, which can help determine its relevance to your inquiry. If the publisher has provided an abstract for the text, you'll see the option to "Show abstract" here. |
| Show related content |
Suggests items that contain similarities in text, which supports discovery and exploration. Note that these may differ from the “Related text” suggestions available under the document viewer, which are based on usage patterns. |
| Recommend topics |
Suggests key concepts that are relevant to the text, which can provide ideas for further exploration. Clicking any of the suggested topics will perform a JSTOR search in a new tab or window. |
| Ask a question |
Type your question in the chat box. The research tool will only answer questions based on the item itself, including its full text and metadata. See Asking Questions About a Text for more information and example questions. |
Using the text
You can learn more about a specific phrase or passage in the text by highlighting it in the document viewer and choosing one of the following options:
- Find related content - Suggests the top 5 items on JSTOR related to the selected text (up to 1,000 characters), with the option to see more results
- What is this selection about? - Suggests the key points of the selected text (1,000 - 6,000 characters)
Downloading your conversation history
You can download a transcript of your current conversation with the research tool as a PDF or HTML file by opening the overflow dropdown menu and selecting Download as PDF or Download as HTML.
For the best results, download your conversation before leaving the item page, since the tool only displays your current conversation for that item for up to 3 hours.
Citing the research tool and responses
In addition to citing content you find on JSTOR, you can also cite the research tool itself or a specific response from the tool in MLA, Chicago, or APA formats. A stable link to the item you're evaluating will be included in the citation.
- To cite the tool in general, open the overflow dropdown menu and select Cite research tool.
- To cite an individual response, use the Cite response button following the response. Your prompt will be included in the citation.
- If you copy a response or part of a response from the tool, a citation will automatically be included when you paste the text.
As with any citation, always review your references and make necessary corrections. Also note that citations for AI tools are still evolving, so follow your discipline or institution-specific guidelines.