AI Tools for Research

General GenAI tools like ChatGPT can be powerful tools for getting started and helping you with your research. It's great for:

• Brainstorming research questions, themes, subtopics, and keywords to use in traditional databases
• Finding relevant journals, databases and researchers
• Getting quick summaries (especially for existing documents, well-defined and represented topics on the internet, and areas you already know something about to make fact checking and spotting inconsistencies easier)
• Translating materials and more

However it does have its pitfalls, as you may have seen with made up citations and incorrect facts. According to Carroll and Borycz (2024), GenAI for research is inconsistent (it won't give you the exact same output twice and is prone to hallucination), incoherent (it often won't give you correct citations for its output), and incomplete (its information is limited).

For more information on using general GenAI tools more effectively in your research see the resources below.

Carroll, A. J., & Borycz, J. (2024). Integrating large language models and generative artificial intelligence tools into information literacy instruction. The Journal of Academic Librarianship, 50(4), 102899. https://doi.org/10.1016/j.acalib.2024.102899

In 2025, updated general GenAI models from ChatGPT, Gemini, Perplexity, and Claude now often include Deep Research features that use newer "reasoning" models that are touted as being a step closer to "agentic AI" - a phrase that's fuzzy but often means AI that can work more autonomously.

-from Wonder Tools

These Deep Research models, often billed as "comprehensive" research assistants, are better at:

• step by step prompting themselves
• showing thier work
• staying on task 
• using retrieval augmented generation and citing specific sources
• using models that are more suited to complex tasks rather than simple queries and
• producing long form results.

However, please be aware that they:

• often require the paid tier of GenAI products
• still do hallucinate and have the usual GenAI pitfalls
• are based on the internet and don't have access to the paywalled literature in our subscription databases and so it's really important to vet the sources and also use traditional methods
• are much slower and more resource intensive
• still benefit from thoughtful prompting

For more information, see some of the resources below and please note this is a rapidly evolving space.