📚

AI Citation Hallucination Checker

Spot fake citations from ChatGPT/Claude in seconds

📚
Learn more — how it works, FAQ & guide
Click to expand

AI Citation Hallucination Checker

Paste citations from ChatGPT/Claude output. Heuristic flags for: implausible DOIs, suspicious year-author combinations, fabricated journal names, unrealistic page numbers. Manual verification still required.

How to use this tool

  1. 1

    Paste citations

    One per line, any format.

  2. 2

    Heuristic flags

    DOI plausibility, year-author combos.

  3. 3

    Manual verify

    Each citation needs Google Scholar check.

Frequently Asked Questions

How big is the citation hallucination problem?
Studies of GPT-3.5: 47% of medical citations were fabricated. GPT-4: ~18%. Even when names + years are real, DOIs + journal names are often wrong. Critical risk for students/researchers.
Can ANY citation from AI be trusted?
Never without verification. Even when "real", the AI may have misattributed findings or cited the wrong paper for a claim. Always Google Scholar / DOI-resolve every citation.
Best workflow?
1) Generate cite with AI, 2) Search Google Scholar by author+year, 3) Verify the cited claim is in the actual paper, 4) Always include DOI in final write-up.

You might also like

🔒
100% Privacy. This tool runs entirely in your browser. Your data is never uploaded to any server.