
A literature review is one of the most time-consuming parts of research. It involves searching for studies, reading and summarising them, comparing findings, and identifying gaps. It is no surprise that many researchers and students have turned to ChatGPT and similar AI tools for help. But how useful are these tools for literature reviews, and what are the risks?
This guide to ChatGPT for literature reviews explains what AI can and cannot do, practical ways to use it responsibly, the dangers of fabricated references, how to verify outputs, and how to disclose AI use.
Table of Contents
What Is a Literature Review?
A literature review summarises and analyses existing research on a topic. Depending on its purpose, it may be:
- Narrative review: a broad overview of a field
- Scoping review: mapping the range of available evidence
- Systematic review: a structured, reproducible search and analysis following strict methods
- Thesis literature chapter: positioning new research within existing knowledge
Each type requires accurate sources, careful reading, and critical thinking.
What ChatGPT Can Help With
| Task | How AI Can Help | Caution |
|---|---|---|
| Brainstorming | Suggest themes, questions, and keywords | Ideas must be checked against real literature |
| Search terms | Propose synonyms and search strings | Test them in proper databases |
| Explaining concepts | Simplify unfamiliar terms or methods | Explanations may contain errors |
| Structuring | Suggest outlines and section headings | Adapt to your actual findings |
| Language editing | Improve clarity and grammar | Check meaning has not changed |
What ChatGPT Should Not Be Used For
- Finding sources without verification: general chatbots can invent references
- Replacing reading: AI summaries cannot substitute for reading the original papers
- Making evidence judgements: assessing study quality requires expert critical appraisal
- Writing the review for you: this may breach academic integrity rules
- Handling confidential material: unpublished manuscripts or sensitive data should not be entered into public tools
The Problem of Fabricated References
One of the biggest risks is that AI chatbots may produce citations that look real but do not exist, or attach real authors and journals to papers they never published. These are often called hallucinated citations.
Fabricated references have caused problems in academic papers, student assignments, and even court cases. Including them in a literature review can lead to rejected manuscripts, failing grades, or misconduct findings.
Why This Happens
Large language models generate text by predicting likely word patterns. They know what a citation usually looks like, so they can produce convincing references even when they lack real information.
How to Use ChatGPT Responsibly in a Literature Review
Step 1: Define Your Question
Use AI to help refine a broad topic into a focused research question, then confirm it with your supervisor or team.
Step 2: Build Search Strategies
Ask for keyword ideas, synonyms, and possible search strings. Then run searches in trusted databases relevant to your field, such as academic library catalogues and discipline-specific indexes.
Step 3: Find Real Sources Yourself
Locate papers through databases, library tools, and reference lists of key articles. Download and read the full texts.
Step 4: Use AI to Aid Understanding
If a method or concept is unclear, AI can offer a simple explanation. Always confirm explanations against textbooks or the original paper.
Step 5: Organise Your Notes
AI may help suggest ways to group themes or compare findings, but your notes should be based on your own reading.
Step 6: Write and Edit
Write the review in your own words. AI can help improve clarity, but you remain responsible for accuracy and interpretation.
Step 7: Verify Everything
Check every citation, quotation, statistic, and claim against original sources before submission.
AI Tools Designed for Research
Some AI research tools are connected to academic databases and are designed to retrieve and summarise real papers. These can be more useful for literature searching than general chatbots, but they still make mistakes, may miss important studies, and can misrepresent findings. They should support, not replace, systematic searching and careful reading.
Special Rules for Systematic Reviews
Systematic reviews require transparent, reproducible methods. If AI is used for tasks such as screening titles and abstracts, researchers should:
- Describe the tool and its role in the methods section
- Have human reviewers check AI decisions
- Report any validation of the tool’s accuracy
- Follow relevant reporting guidelines and journal policies
Disclosing AI Use
Most journals and many universities require disclosure of substantive AI use. A disclosure should name the tool, explain how it was used, and confirm that the authors verified the content and take full responsibility. AI tools should never be listed as authors.
Example Prompts That Work Well
Prompts that ask AI to support your thinking, rather than replace your research, tend to be safer and more useful:
- “Suggest synonyms and related terms for searching studies on [topic].”
- “Explain the difference between a cohort study and a case-control study in simple terms.”
- “Propose a possible outline for a literature review on [topic] with five themes.”
- “Rewrite this paragraph I wrote to make it clearer without changing its meaning.”
- “What questions should I ask when critically appraising a randomised trial?”
Avoid prompts such as “Give me ten references on this topic” or “Write my literature review,” which are more likely to produce fabricated sources or breach integrity rules.
How to Check a Reference Quickly
- Search the exact title in a trusted academic database or library catalogue
- Confirm the authors, journal, year, volume, and pages match
- Look up the DOI and check that it leads to the correct article
- Open the full text and confirm it says what you claim
- Be suspicious if a reference cannot be found anywhere
Benefits for Non-Native English Writers
For researchers writing in a second language, AI tools can make it easier to express ideas clearly and read complex papers. Used carefully, this can support fairer participation in international research. Writers should still review all changes and ensure that the meaning reflects their own understanding.
Common Mistakes to Avoid
- Copying AI-generated references into your reference list
- Relying on AI summaries without reading papers
- Ignoring your institution’s AI rules
- Entering confidential or unpublished data into public tools
- Failing to disclose AI use
Keep a Record of AI Use
Noting which tools you used, when, and for what tasks makes disclosure simple and helps you answer questions from supervisors, examiners, or editors. It also reminds you which parts of your work need extra checking before submission.
Working With Your Supervisor
Students should discuss AI use with their supervisors before starting a literature review. Agreeing on acceptable uses, disclosure requirements, and checking methods early prevents misunderstandings later and ensures the review meets academic expectations.
Final Thoughts
ChatGPT for literature reviews can be a helpful assistant for brainstorming, building search terms, explaining concepts, and improving writing. But it is not a reliable source of references or a substitute for reading and critical thinking. By finding sources yourself, verifying every citation, following institutional rules, and disclosing AI use, you can benefit from AI while keeping your literature review accurate and trustworthy.





