User Intent to Use DeepSeek for Health Care Purposes and Their Trust in the Large Language Model: Multinational Survey Study
Bibliographic record
Abstract
BACKGROUND: Generative artificial intelligence (Gen-AI)-particularly large language models (LLMs)-has generated unprecedented interest in applications ranging from everyday Q&A to health-related inquiries. However, little is known about how everyday users decide whether to trust and adopt these technologies-particularly in high-stakes contexts like personal health. OBJECTIVE: This study examines how ease of use, perceived usefulness, and risk perception interact to shape user trust in and intentions to adopt DeepSeek, an emerging LLM-based platform, for healthcare purposes. METHODS: We adapted survey items from validated technology acceptance scales to assess user perception of DeepSeek, focusing on constructs such as trust, intent to use for health, ease of use, perceived usefulness, and risk perception. A 12-item Likert scale questionnaire was developed and pilot-tested (n=20) for clarity and consistency. It was then distributed online to users in India (IND), United Kingdom (UK), and United States of America (USA) who had used DeepSeek within the past two weeks. Data analysis involved descriptive frequency assessments and Partial Least Squares Structural Equation Modeling (PLS-SEM) to evaluate the measurement and structural models. Structural equation modeling assessed direct and indirect effects, including potential quadratic relationships. RESULTS: A total of 556 complete responses were collected, with respondents almost evenly split across IND (n=184), the UK (n=185), and the USA (n=187). Regarding AI in healthcare, when asked if they were comfortable with their healthcare provider using AI tools, 59.3% (n=330) were fine with AI use provided their doctor verified its output, and 31.5% (n=175) were enthusiastic about its use without conditions. DeepSeek was used primarily for academic and educational purposes, 50.7% (n=282) used DeepSeek as a search engine, and 47.7% (n=265) for health-related queries. When asked about their intent to adopt DeepSeek over other LLMs like ChatGPT, 52.1% (n=290) were likely to switch, and 28.9% (n=161) were very likely to do so. The study revealed that trust plays a pivotal mediating role: ease of use exerts a significant indirect impact on usage intentions through trust. At the same time, perceived usefulness contributes to trust development and direct adoption. By contrast, risk perception negatively affects usage intent, emphasizing the importance of robust data governance and transparency. Significant non-linear paths were observed for ease of use and risk, indicating threshold or plateau effects. CONCLUSIONS: Users are receptive to DeepSeek when it's easy to use, useful, and trustworthy. The model highlights trust as a mediator and shows non-linear dynamics shaping AI-driven healthcare tool adoption. Expanding the model with mediators like privacy and cultural differences could provide deeper insights. Longitudinal or experimental designs could establish causality and track user attitudes. Further investigation into threshold and plateau phenomena could refine our understanding of user perceptions as they become more familiar with AI-driven healthcare tools.
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.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.005 | 0.014 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.001 |
| Bibliometrics | 0.001 | 0.001 |
| Science and technology studies | 0.001 | 0.001 |
| Scholarly communication | 0.001 | 0.002 |
| Open science | 0.000 | 0.001 |
| Research integrity | 0.001 | 0.001 |
| Insufficient payload (model declined to judge) | 0.002 | 0.001 |
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 itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".