Evaluating Harm Reduction Advice From Large Language Models
Khy Choo, Nathan Smith, Bernadette Ward, Robert M. Lundin · 2026
Read Evaluating Harm Reduction Advice From Large Language Models by Khy Choo, Nathan Smith, Bernadette Ward, Robert M. Lundin (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/PMC13631077.txt.
Description
Khy Choo and three colleagues tested how well four popular chatbots answer harm reduction questions from people who inject drugs. Their focus was practical: how to filter dissolved pills through a wheel filter before injecting, and where to find local services. They built a scoring rubric and graded each model's answers.
The gap was wide. Two systems scored high, giving thorough filter instructions and accurate local service details, while the other two scored poorly, omitting the filter guidance entirely and answering thinly elsewhere. Some responses also contained misinformation, and built-in safety limits blocked some questions outright. The authors see chatbots as a possible supplement to specialist services, never a substitute, and offer their rubric for reuse.
Catalog record
| CREATOR | Khy Choo, Nathan Smith, Bernadette Ward, Robert M. Lundin |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 151 KB |
| LICENSE | https://creativecommons.org/licenses/by/4.0 |
| RIGHTS | CC BY |
| SOURCE | Europe PMC |