Use of Artificial Intelligence for Medical Literature Search: Randomized Controlled Trial Using the Hackathon Format
Bibliographic record
Abstract
BACKGROUND: Mapping out the research landscape around a project is often time consuming and difficult. OBJECTIVE: This study evaluates a commercial artificial intelligence (AI) search engine (IRIS.AI) for its applicability in an automated literature search on a specific medical topic. METHODS: To evaluate the AI search engine in a standardized manner, the concept of a science hackathon was applied. Three groups of researchers were tasked with performing a literature search on a clearly defined scientific project. All participants had a high level of expertise for this specific field of research. Two groups were given access to the AI search engine IRIS.AI. All groups were given the same amount of time for their search and were instructed to document their results. Search results were summarized and ranked according to a predetermined scoring system. RESULTS: The final scoring awarded 49 and 39 points out of 60 to AI groups 1 and 2, respectively, and the control group received 46 points. A total of 20 scientific studies with high relevance were identified, and 5 highly relevant studies ("spot on") were reported by each group. CONCLUSIONS: AI technology is a promising approach to facilitate literature searches and the management of medical libraries. In this study, however, the application of AI technology lead to a more focused literature search without a significant improvement in the number of results.
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 | 0.022 | 0.037 |
| Meta-epidemiology (narrow) | 0.002 | 0.001 |
| Meta-epidemiology (broad) | 0.006 | 0.004 |
| Bibliometrics | 0.003 | 0.003 |
| Science and technology studies | 0.001 | 0.003 |
| Scholarly communication | 0.003 | 0.003 |
| Open science | 0.002 | 0.002 |
| Research integrity | 0.004 | 0.003 |
| Insufficient payload (model declined to judge) | 0.016 | 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".