Using generative AI chatbots such as ChatGPT, Google Gemini, etc. to find academic sources can be an ineffective strategy due to several key limitations:
1.) Hallucinations
AI chatbots frequently provide incorrect information - called "hallucinations"
In academic research, hallucinations most commonly appear as citations to articles that don't exist
These fake citations can seem convincing because they often include:
Real authors' names
Actual journal titles
Plausible-sounding article titles
Hallucinations are not a temporary bug that will eventually be fixed. Recent research from OpenAI (the organization behind ChatGPT) has found that hallucinations are an inevitable feature of current chatbot training methods
2.) Information Quality
Even when chatbots cite real articles, their relevance and quality aren't guaranteed. Just because an AI tool provides an answer does not mean it will be good or relevant. Critical evaluation is always necessary
AI chatbots operate with significant unknowns:
What sources are in their training data?
How current is their information? (Most have knowledge cutoffs)
What criteria determine which sources they prioritize?
Are they accessing paywalled databases or only free web content?
3.) Limited Search Control
Unlike library databases, you cannot easily filter or refine chatbot results using:
Date ranges
Peer-review status
Subject headings
Other standardized search filters
Chatbot responses are not reproducible - using the same exact prompt will generate similar but not identical results each time, making it difficult to revisit earlier searches
For these reasons, the library's academic databases remain the more reliable and efficient tools for finding scholarly resources.
While not very effective for finding academic sources, AI chatbots can be helpful with preliminary research tasks:
Brainstorming research topics, ideas, or keywords
Getting quick overviews of unfamiliar concepts or terminology
Summarizing complex articles you've already found and verified through library databases
Better approach: Use AI tools embedded within library databases. These specialized tools:
Are designed for specific research functions
Have narrower focus that reduces hallucination risk
Provide more accurate, reliable results
The library provides AI-powered research tools that are more reliable than general chatbots because they work with verified academic content:
Natural Language Search (QuickSearch) - Search using natural language questions
"More Like This" (databases & QuickSearch) - Find similar articles automatically
Scopus AI - Generate summaries from peer-reviewed literature
These tools are designed specifically for academic research and have lower risk of hallucinations.
Learn how to use these tools: See the AI Tools in the SHU Library page in the AI & Academic Research Guide for detailed instructions.
Always check your instructor's AI policy before using any AI tools in your coursework. When in doubt, be sure to ask!

FIND US ON