Best Practices for Using AI on Research Papers
Schools, publishers, and academic associations are developing policies for the use of AI in academic work, but in nearly every case, the power to accept or reject AI use in course assignments lies with individual instructors. Before you turn in work that was created with the help of AI, check the policies of your teacher, advisor, or publisher.
To that end, keep careful track of any AI platforms you use during your research process, including any prompts, and the results yielded. This is crucial to your being able to accurately acknowledge and cite your AI-assisted work.
Managing and citing your uses of AI and other resources helps keep your academic writing accurate and fair; moreover, acknowledging resources strengthens the scholarly community as resources develop and change the processes of work in the field.
AI Can Create Concept Maps
Mind-mapping is a visualization technique that can help with topic and thesis development. Creating a mind map generally starts with the writer stating a concept or theme and building out a variety of related and new ideas from that original theme. In the process of building this structure of ideas, the writer forms connections and hierarchies themes, subthemes, topics, subtopics, and examples.
Creating a mind map can be particularly helpful when writers have a complex concept that could be developed and organized along several different paths or when they have a long project like a capstone or dissertation.
AI tools for mind-mapping can help with visualizing concepts and showing connections among topics and subtopics. Below are two AI-supported mind-mapping tools available to researchers at the University of Bridgeport.
Develop Topics with AI
Several AI systems provide tools that can be used for academic topic development, even though several popular programs were developed for general use. For academic use, verify sources it suggests by searching for them in the University of Bridgeport library databases
Asking Questions With AI to Generate and Clarify Topics, Search
ChatGPT from OpenAI answers questions and generates ideas.
Developed by Anthropic AI, Claude AI generates conversations, summary and some analysis.
Copilot searches and generates images and text.
Answers questions and generates topics
Gemini is Google's AI for answering questions and searching general information.
AI tool for editing and submission-readiness checks on papers.
Natural language searching and conversation, text summary, follow-up questioning.
AI tool for article discovery.
AI Tools for Conducting and Managing Research
Consensus is a tool that uses prompts to find, synthesize, and analyze relevant sources in science topics, providing 10 summaries and references for papers. Connects to pdfs of some papers, but otherwise students should find the actual paper through the UB library and databases. Librarians' reviews of Consensus suggest that the system can save time for literature reviews researchers should use caution in relying on the conclusions the AI presents about individual papers.
PDFgear is designed to assist you with both summarizing and taking a deep dive into a paper that you provide. Once you upload a resource, PDFgear will create a pdf reader for it that is interactive and will answer your questions and show you where in the paper the answer comes from. this is a good, free alternative to PDFChat. PDFgear Copilot features natural language processing.
Scispace is a finding tool that answers questions in the sciences and cites the resources containing these answers. Scispace can also provide a summary and a breakdown of the paper sections. Librarians' reviews suggest Scispace can be a useful timesaver in aggregating and organizing research but that its data and research conclusions should not be relied upon at face value.
Research Rabbit combines elements of visualization/concept mapping tools with finding tools for research papers. Its searches are powered by PubMed for life sciences and biomedical fields and by Semantic Scholar for all other fields. As you choose papers to add to your collection, Research Rabbit will find others with relevant keywords. Some librarian reviews suggest the data are not up-to-date and should be verified.
Perplexity is another search tool. It draws from a very wide variety of sources, which allows its responses to be up-to-date but which also means that some information is from less academically credible sources than others. However, the AI provides the sources it has used in answering the question, which can help in judging and verifying. Perplexity also searches images. Ultimately, the user should verify Perplexity's information and language rather than merely cut/paste/turn-in.
Semantic Scholar specializes in academic literature and is an AI search tool along the lines of Google Scholar. It features search results that are not behind a paywall. Some reviewers suggest users should verify all summary and paraphrase for accurate reporting of articles.
Quillbot is an AI tool that creates paraphrases and summaries even while it checks your paper for grammar and detects plagiarism and AI. UB librarians' tests of Quillbot found that here, as with Grammarly and other such platforms, the writer should remember to revise and rewrite carefully rather than rely on copying the exact language of the paraphrases and summaries, which do not always adhere to the standards of academic writing and acceptable paraphrase/summary (both of which must be accurately cited in every case).
Learn more about these AI models.
Artificial Intelligence (AI) in Research © 2025 by Mary Lamothe, PhD is licensed under CC BY 4.0 
