Health Consumers' Use and Perceptions of Health Information from Generative Artificial Intelligence Chatbots: A Scoping Review
Why this work is in the frame
A frame that forgets how it found something cannot be audited. These are the routes that admitted this work.
Bibliographic record
Abstract
Abstract Health consumers can use generative artificial intelligence (GenAI) chatbots to seek health information. As GenAI chatbots continue to improve and be adopted, it is crucial to examine how health information generated by such tools is used and perceived by health consumers. To conduct a scoping review of health consumers' use and perceptions of health information from GenAI chatbots. Arksey and O'Malley's five-step protocol was used to guide the scoping review. Following PRISMA guidelines, relevant empirical papers published on or after January 1, 2019, were retrieved between February and July 2024. Thematic and content analyses were performed. We retrieved 3,840 titles and reviewed 12 papers that included 13 studies (quantitative = 5, qualitative = 4, and mixed = 4). ChatGPT was used in 11 studies, while two studies used GPT-3. Most were conducted in the United States (n = 4). The studies involve general and specific (e.g., medical imaging, psychological health, and vaccination) health topics. One study explicitly used a theory. Eight studies were rated with excellent quality. Studies were categorized as user experience studies (n = 4), consumer surveys (n = 1), and evaluation studies (n = 8). Five studies examined health consumers' use of health information from GenAI chatbots. Perceptions focused on: (1) accuracy, reliability, or quality; (2) readability; (3) trust or trustworthiness; (4) privacy, confidentiality, security, or safety; (5) usefulness; (6) accessibility; (7) emotional appeal; (8) attitude; and (9) effectiveness. Although health consumers can use GenAI chatbots to obtain accessible, readable, and useful health information, negative perceptions of their accuracy, trustworthiness, effectiveness, and safety serve as barriers that must be addressed to mitigate health-related risks, improve health beliefs, and achieve positive health outcomes. More theory-based studies are needed to better understand how exposure to health information from GenAI chatbots affects health beliefs and outcomes.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
Full frame distilled prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. Learned from the 10,348 direct Codex labels and 10,348 direct Gemma labels. Candidate is the union of thresholded teacher heads; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels or direct frontier model labels.
Codex and Gemma teacher scores by category
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.012 | 0.004 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.005 | 0.000 |
| Bibliometrics | 0.000 | 0.001 |
| Science and technology studies | 0.001 | 0.000 |
| Scholarly communication | 0.000 | 0.002 |
| Open science | 0.000 | 0.000 |
| Research integrity | 0.001 | 0.002 |
| Insufficient payload (model declined to judge) | 0.000 | 0.000 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from it