JSTOR Introduces AI-Assisted Dynamic Search for Research Discovery

7 September 2026

JSTOR has begun rolling out a new AI-assisted dynamic search that gives researchers another way to discover academic literature by exploring the meaning and context of a research question rather than relying entirely on exact keywords.

 

The feature is being introduced gradually from 7 October 2026 at participating institutions where JSTOR’s AI-enabled features are available. It will work alongside the platform’s traditional keyword search, giving researchers an additional option when beginning a literature search, exploring a new subject or working with terminology they may not yet know well.

 

Traditional keyword searching is particularly useful when a researcher already knows the words, phrases, authors or concepts likely to appear in relevant literature. Dynamic search approaches the task differently by considering the meaning behind a question and identifying sources that discuss related ideas, even when those sources use different terminology.

 

A researcher studying ecological resilience in urban planning, for example, may also encounter literature on climate adaptation, sustainable cities or urban green infrastructure. This can help researchers discover connections between topics that may be difficult to identify through exact keyword combinations alone.

 

According to JSTOR, dynamic search combines several search approaches to identify and rank material that is conceptually related to a researcher’s question. It can also suggest related searches that allow users to broaden, narrow or explore different directions within a topic.

 

The search results continue to connect researchers directly to the original scholarly content available on JSTOR. Researchers can then read the material, assess its relevance, examine the evidence and determine whether it supports their work.

 

This remains an important part of JSTOR’s approach to artificial intelligence. The platform describes its AI features as tools that support research and discovery rather than systems that replace the researcher’s judgement. Reading, evaluating sources, interpreting evidence and drawing conclusions remain the responsibility of the researcher.

 

The new feature complements JSTOR’s existing AI research tool, which helps users explore material they are reading, identify important ideas and investigate related topics. Together, the tools are intended to make it easier for researchers to navigate large collections while maintaining direct access to the original sources.

 

Dynamic search currently focuses on journal articles, book chapters and research reports. Researchers who need exact phrases, Boolean operators, field-specific searching or access to other types of material can continue using JSTOR’s keyword search.

 

Institutional control remains part of the rollout. Universities and other participating institutions decide whether JSTOR’s on-platform AI features are available to their users. Institutions that previously opted out of or disabled JSTOR’s AI research tool will have dynamic search disabled by default.

 

Researchers at institutions where the feature is enabled will also need to access JSTOR through their institution and sign in to a personal JSTOR account before the dynamic search option becomes available.

 

The introduction of dynamic search reflects JSTOR’s continuing effort to explore how artificial intelligence can support academic research without separating researchers from the scholarly sources on which their work depends. By combining traditional keyword searching with concept-based discovery, the platform is giving researchers another way to identify relevant literature and explore unfamiliar areas of study.

 

Researchers and librarians can learn more about the feature through JSTOR Support, while information about institutional access to AI-enabled features is available through JSTOR’s guidance for administrators.

 

More information about JSTOR’s wider approach to artificial intelligence is also available on its AI at JSTOR page.