Reliability and readability of five AI chatbots for concussion health advice across retrieval augmented and pretrained models
Hefang Huang, Caihong Zhang, Hao Hu, Xiang Tong · 2026
Read Reliability and readability of five AI chatbots for concussion health advice across retrieval augmented and pretrained models by Hefang Huang, Caihong Zhang, Hao Hu, Xiang Tong (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/pmc-13328287.txt.
Description
Hefang Huang, Caihong Zhang, Hao Hu, and Xiang Tong audit five AI chatbots giving advice on concussion, checking both reliability and readability. Patients increasingly ask chatbots about head injuries, and wrong or unreadable advice is a safety risk. Their 2026 cross-sectional audit used eleven common patient questions drawn from search trends, asked fresh with no priming. Two blinded neurosurgeons scored the answers against the 2023 Amsterdam consensus statement, using standard quality instruments. Retrieval-augmented models were compared with plain pretrained ones.
The design separates architecture from brand: does bolting on retrieval actually help? Blinded expert scoring against a consensus text keeps the verdict honest. The authors supply a framework others can rerun as models change. Their concussion focus is deliberate: advice errors there can send a patient back onto the field. Readability counts too, since perfect advice nobody understands helps no one.
Catalog record
| CREATOR | Hefang Huang, Caihong Zhang, Hao Hu, Xiang Tong |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 40 KB |
| VERSION | 1.0 |
| LICENSE | CC BY 4.0 |
| RIGHTS | CC BY |
| SOURCE | Europe PMC |