Research and Creation with Generative AI Tools
In addition to the usual considerations for using Generative AI tools, for use in research, see this presentation for more considerations specific to research. In general, remember that these tools are often a black box and likely biased. Large language models (LLMs) also are known to “hallucinate” (e.g. make things up); they are designed to give you the statistically most likely next words or output - prioritizing fluency over accuracy (GPT 4o is better but not perfect). Research tools based on the research literature are more likely to use retrieval augmented generation, a process by which the chatbot bases its response on a defined set of papers (with actually existing references, unlike some of the more general chatbots like ChatGPT); this can reduce but doesn't eliminate hallucinations. And none of these chatbots know anything about the real world. Remember, these are “stochastic parrots” (Bender et. al., 2021), not all-knowing magic beings or anything like a human intelligence no matter how fluent they are!
Generative AI Research Tools
In addition to general generative AI tools (probably based on data from the Internet amongst other sources) like ChatGPT (and products based on it like SearchGPT), Microsoft Copilot, Google's Gemini (and products based on it like NotebookLM and DeepMind), Claude, Perplexity (and Perplexity Spaces) and others mentioned on the Generative AI guide, there are tools specifically for research, often based on the research literature, that incorporate generative AI. Here are just a few below. This space for higher education is changing rapidly so keep an eye on Ithaka S&R's Generative AI Product Tracker and check out Other Tools to Investigate below.
Wondering which tools to use when? Consider your workflow (and whether you need GenAI at all), the task you are trying to accomplish, the level of accuracy and privacy you need, the policies for whatever task you're doing (for example, does your professor or the journal you're submitting to or your work have an AI policy). You might also want to look at librarian Nicole Hennig's guide, "Which generative AI tool for your task?".
Note: Many of these tools are experimental and are in beta (or alpha) so consider them pilots, not the last word for comprehensive searching and research. Nearly all of these are freely available or have a free tier. For more information on some of these see this presentation. Also note that these lists for the most part don't include AI tools or capabilities that are built into or add ons to our existing subscription databases (such as Scopus AI, Science Direct AI, Primo Research Assistant, JSTOR Interactive Research Assistant, etc - which all require additional subsciptions fees that we have not added at this time).
Scite This link opens in a new window"Scite is an award-winning platform for discovering and evaluating scientific articles via Smart Citations. Smart Citations allow users to see how a publication has been cited by providing the context of the citation and a classification describing whether it provides supporting or contrasting evidence for the cited claim." Scite also does reference checking, can create custom dashboards and alerts, and more. Create an account with your W&M email to access all features. Scite has a free browser plug in and also integrates into Zotero. Scite is also trying to integrate chatbot features based on the literature, not the internet (Scite Assistant - in beta) so it can provide correct citations though always double check that they say what the citation says they are saying. Also it allows you to adjust Assistant responses. Find tutorials on YouTube: https://www.youtube.com/channel/UCso9-KdHZ6vfbYgM59zp9EA and a recent W&M presentation at https://docs.google.com/presentation/d/10qBk49Ytw--id8nKJvOAfot0V_rIdzyESvHxqicAX98/edit?usp=sharing
Consensus
Question-based search - ask a question to mine answers from research papers. Focused on health but expanding to questions in other areas (with caveats about accuracy). See Consensus’ best practices: https://consensus.app/home/blog/maximize-your-consensus-experience-with-these-best-practices/. Tries to pull out quotes and synthesize
Elicit
"Elicit is a research assistant using language models like GPT-3 to automate parts of researchers’ workflows. Currently, the main workflow in Elicit is Literature Review. If you ask a question, Elicit will show relevant papers and summaries of key information about those papers in an easy-to-use table." Uses elements of earlier versions of GPT as well as other tools to find a smaller selection of relevant papers to answer a question and summarizes them. Often health-care related; may want to use to build a broad bibliography in heath areas. Built by Ought, a non-profit machine learning research lab using the Semantic Scholar database. Initially gives you eight papers (not only based on keyword matching but also semantic similarity) and summarizes them (creates its own abstracts). Helps find initial leads, answer questions, find research directions. Can search by study type and add information like interventions and outcomes. Not for comprehensive search and favors empirical scientific and social scientific topics.
SciSpaceFree. Has browser extension for Chrome.
NotebookLMGoogle's beta AI research assistant
Apps
GPTs (apps built on top of ChatGPT) for research includes ones such as Scholar AI, Scholar GPT, Research Papers, and more
Other AI Discovery Tools
Not all AI research tools are generative AI tools like ChatGPT. Many of the tools below may be using some combination of machine learning, deep learning, natural language processing and more to find and visualize the research literature. Some, but not all, of these also incorporate generative AI.
Scite This link opens in a new windowScite is an award-winning platform for discovering and evaluating scientific articles via Smart Citations. Smart Citations allow users to see how a publication has been cited by providing the context of the citation and a classification describing whether it provides supporting or contrasting evidence for the cited claim.
Research Rabbit
Research discovery app. Uses seed papers to suggest possibly relevant papers. Can also connect Zotero and can visualize networks as well as trace citation chains and similar articles and authors.
Connected Papers
Explore connected papers in a visual graph by entering a paper identifier. Useful for producing visualizations of networks of papers.
Dimensions
"Dimensions is a linked research knowledge system that re-imagines discovery and access to research. Developed by Digital Science in collaboration with over 100 leading research organizations around the world, Dimensions brings together grants, publications, citations, alternative metrics, clinical trials, patents and policy documents to deliver a platform that enables users to find and access the most relevant information faster, analyze the academic and broader outcomes of research, and gather insights to inform future strategy. For more information about the product have a look at our support site or visit https://dimensions.ai."
Inciteful
Tool for finding literature in two ways - not only useful for analyzing and visualizing networks of literature but also for seeing papers that bridge between two different subject areas (useful for interdisciplinary research).
Keenious
Find research relevant to documents. Uses “seed” document in PDF, link or pasted text. Ranks based on relevance, shared terms, and predicted meaning. Gives related articles. Can filter by date, citation count and keyword and gives topics covered by that article and related ones. Can get add in for Word and Gdocs to have it analyze your own text and suggest articles, even for specific sections. NOTE: save your results (full text PDFs) as you go since there is limited bookmarking and citation export capability.
LitMaps
Literature review software. Search, discover, visualize and analyze the literature.
R Discovery
With tools like RDiscovery you can get literature recommendation sent to a dashboard feed (which also produces email alerts). You give it topics you’re interested in and it will populate a feed with articles it things are relevant (though your mileage may vary).
Semantic Scholar
Science-focused search. Can filter by field/subject, date, availability of PDF, author, and journal or conference. Provides short AI-generated abstract summaries and ability to see highly cited papers quickly. Provides paper recommendations and alerts and has beta research tools.
Lens.org
"Serves linked open knowledge artefacts and metadata with tools to inform effective, efficient and equitable problem solving." Not AI based. Also useful for the patent literature.
Visual Search and Visualization Tools
Other Discovery and Visualization Tools-Uses
Other AI-powered discovery and visualization tools can be:
Useful for:
- Finding relevant papers you might not ordinarily have found by means other than keyword matching, either by searching similar or seeding relevant papers. NOTE: systems that work from seed papers are often better than those that ask you topics you’re interested in or work like recommender systems (e.g. Amazon “you might also like”), though none of them are perfect
- Visualizing networks of citations and researchers
Not recommended for:
- Comprehensive searches or those requiring exact replication. Uses these as a supplement to library databases and Google Scholar
- Humanities and non-empirical social sciences may be much less well represented depending on corpus of text the search is based on and results vary by field and subfield
- Interdisciplinary research (except for Inciteful), especially for recommender-based systems
