Assessment of vaccine information accuracy across large language models
Julia Siebielec, Filip Raciborski · 2026
Read Assessment of vaccine information accuracy across large language models by Julia Siebielec, Filip Raciborski (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/PMC13319027.txt.
Description
Julia Siebielec and Filip Raciborski offer a 2026 assessment of vaccine information accuracy across four large language models. ChatGPT, Copilot, DeepSeek, and Gemini each answered ninety multiple-choice questions drawn from a government vaccine website. Data were collected in March 2026. Two open questions probed sources and the pros and cons of universal vaccination.
Gemini scored highest, with a correct share of 0.98 across the test set. ChatGPT scored lowest at 0.86 in the same test. Results differed significantly between models. On open questions, all named suitable information sources. All cited herd immunity as an advantage and adverse reactions as a disadvantage. Answers were correct but thin. The authors judge the models a limited information source on their own.
Catalog record
| CREATOR | Julia Siebielec, Filip Raciborski |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 49995 |
| LICENSE | https://creativecommons.org/licenses/by/4.0 |
| RIGHTS | CC BY |
| SOURCE | Europe PMC |