Crowdsourcing a Training Dataset of Question-and-Answer Pairs for AI-Enabled Health Information Tools on Sexually Transmitted Infections: Protocol for a Cross-Sectional Exploratory Survey Study
Bibliographic record
Abstract
BACKGROUND: Sexually transmitted infections are a significant public health concern, particularly in sub-Saharan Africa, where their prevalence remains high. Promoting awareness and reducing stigma are essential strategies for addressing this challenge, but those affected often have limited access to accurate and culturally appropriate health information. Therefore, innovative solutions are essential to enhance sexual health literacy and encourage informed health-seeking behaviors. Artificial intelligence (AI)-enabled tools, such as chatbots, have emerged as promising avenues for delivering accurate and accessible health information. However, their potential is constrained by the lack of contextualized datasets, which are crucial for ensuring their effectiveness and relevance to diverse populations. OBJECTIVE: This study aims to develop an open access, contextualized dataset of question-and-answer pairs on sexual health and sexually transmitted infections to support the development and training of digital and AI-enabled health information tools. METHODS: Using a crowdsourcing approach, questions are being collected from participants aged ≥15 years via online platforms, paper-based submissions, and in-person interactions at public events across sub-Saharan Africa. Each question will be anonymized and reviewed by medical professionals who will provide accurate, evidence-based answers. The dataset will then undergo processing, including cleaning and tagging for AI training, ensuring adherence to findability, accessibility, interoperability, and reusability principles. The final dataset will be published as open access. RESULTS: Data collection began on June 12, 2024, and is ongoing. The data collection process was piloted in Kigali, Rwanda, where 132 questions were collected. As of August 2025, the study had collected over 5620 question-and-answer pairs. The collected data are undergoing a simultaneous rigorous data processing phase in collaboration with health workers who provide evidence-based answers to the questions and new questions based on their experience in the clinic. The data cleaning and processing will enhance the utility of the data for AI applications. CONCLUSIONS: The final dataset will be published as open access in 2025, contributing to the development of AI-driven health tools and promoting public health literacy. INTERNATIONAL REGISTERED REPORT IDENTIFIER (IRRID): DERR1-10.2196/70005.
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 | 0.028 | 0.058 |
| Meta-epidemiology (narrow) | 0.001 | 0.001 |
| Meta-epidemiology (broad) | 0.001 | 0.002 |
| Bibliometrics | 0.002 | 0.002 |
| Science and technology studies | 0.004 | 0.002 |
| Scholarly communication | 0.003 | 0.002 |
| Open science | 0.003 | 0.005 |
| Research integrity | 0.003 | 0.004 |
| Insufficient payload (model declined to judge) | 0.044 | 0.017 |
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".