For systematic review references, hallucination rates were 39.6% for GPT-3.5, 28.6% for GPT-4, and 91.4% for Bard; validation remained essential. #SystematicReview #ChatGPT #ReferenceHallucination doi.org/10.2196/53164
https://doi.org/10.2196/53164
doi.org
October 13, 2026 at 8:20 PM
For systematic review references, hallucination rates were 39.6% for GPT-3.5, 28.6% for GPT-4, and 91.4% for Bard; validation remained essential. #SystematicReview #ChatGPT #ReferenceHallucination doi.org/10.2196/53164