Today’s post was written by library student assistant Judite do Bem Sampaio. Judite is an international student from Portugal double-majoring in economics and entrepreneurship.
On October 16, 2025, the William H. Hannon Library hosted a workshop entitled “AI-Generated Tools for Academic Research: Featuring “Connected Papers.” The workshop showcased the way artificial intelligence is transforming how academic works are searched, evaluated, and represented.
The presentation began by contrasting historical modes of academic searching with new AI systems. Academic databases previously performed either Boolean or lexical searching: i.e. searching results relied on finding literal phrases or words. The presenter explained that such a constricted approach could ignore existing studies using a different terminology. In contrast, contemporary AI-based tools use semantic and natural language searching, allowing users to find information by meaning and not only word-for-word. These new tools are shaping a more natural, integrated, and insightful research process.
The session identified several key tools that are spearheading this change. OpenAlex was launched as an open-source, free index of scholarly publications aimed at open access and transparency. Semantic Scholar is an AI-powered search engine that uses machine learning to identify conceptual connections between papers rather than being mere functions of citation counts. Crossref was also noted as essential infrastructure for scholarly integrity as it offers stable Digital Object Identifiers (DOIs) that make enduring links between scholarship outputs possible. Together, these sites provide the foundation for a modern, AI-facilitated research ecosystem.
Most of the conversation was about citation-based literature mapping tools, including Connected Papers, Litmaps, and ResearchRabbit. These platforms use citations and algorithms to map the relationships between research studies and return insights to researchers about how ideas evolve over time. The concept of backward and forward citation searching was explained: backward citations identify prior work that a paper builds on, and forward citations identify newer papers that reference the original. Using this approach, researchers can follow the scholarly conversation historically and forward.
The presenter continued on to explain in more detail how Connected Papers works. Powered by Semantic Scholar data, the software builds a force-directed graph: a visual map that clusters papers based on similarity. Each node in the graph is a research paper, with color representing more recent publication dates and node size representing the number of citations. Relative line thickness between nodes indicates strength of similarity between papers. The system reads around 50,000 papers per field to display the 40-50 most applicable to a chosen “seed” paper, allowing users to browse related research in a natural way.
Connected Papers makes use of two measures of similarity, co-citation and bibliographic coupling, to measure these relationships. Co-citation occurs when two papers are cited together by other authors, while bibliographic coupling links papers that cite the same earlier works.
Combining these tools yields an even-handed view of how research questions are connected over time. The lecture highlighted that these devices surpass mere search by giving a visual and mental feel for how information is structured, and therefore they are useful for literature reviews, thesis writing, and detecting research holes.
The talk also covered key features of Connected Papers including integration with reference managers like Zotero in BibTeX format, keyword or year of publication filtering, and search by antecedent works (early studies) or derivative works (follow-up studies based on antecedent ones). A demo based on breast cancer studies illustrated the connection of antecedent works and the following influential works, allowing immediate comprehension of complex citation networks.
Finally, the session discussed ethical standards in the TAAP approach: Transparency and Disclosure, Authority Assessment, Validation for Accuracy, and Privacy Issues. These stand as guidelines that call on AI used in research to be responsible, ensuring tools are accurate, transparent, and considerate of data privacy.
Overall, the presentation showed how AI is revolutionizing scholarly search. From static keyword returns to intelligent, meaning-based and visually focused systems, technologies like Connected Papers and Semantic Scholar are making research more streamlined, more connected, and more insightful. This shift reflects a broader movement toward AI-facilitated discovery: discovery that liberates researchers to focus on creativity and critical thinking rather than painful information retrieval.
What a great chance to learn about how AI is transforming research and helping us discover knowledge in smarter, more interconnected ways!