Using artificial intelligence tools to automate data extraction for living evidence syntheses
Bibliographic record
Abstract
Living evidence synthesis (LES) involves repeatedly updating a systematic review or meta-analysis at regular intervals to incorporate new evidence into the summary results. It requires a considerable amount of human time investment in the article search, collection, and data extraction phases. Tools exist to automate the retrieval of relevant journal articles, but pulling data out of those articles is currently still a manual process. In this article, we present a proof-of-concept Python program that leverages artificial intelligence (AI) tools (specifically, ChatGPT) to parse a batch of journal articles and extract relevant results, greatly reducing the human time investment in this action without compromising on accuracy. Our program is tested on a set of journal articles that estimate the mean incubation period for COVID-19, an epidemiological parameter of importance for mathematical modelling. We also discuss important limitations related to the total amount of information and rate at which that information can be sent to the AI engine. This work contributes to the ongoing discussion about the use of AI and the role such tools can have in scientific research.
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 | 0.132 | 0.330 |
| Meta-epidemiology (narrow) | 0.003 | 0.003 |
| Meta-epidemiology (broad) | 0.003 | 0.007 |
| Bibliometrics | 0.015 | 0.010 |
| Science and technology studies | 0.003 | 0.002 |
| Scholarly communication | 0.012 | 0.007 |
| Open science | 0.004 | 0.008 |
| Research integrity | 0.002 | 0.005 |
| Insufficient payload (model declined to judge) | 0.036 | 0.010 |
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".