Development of an Evaluation Instrument on Artificial Intelligence Search Tools for Evidence Synthesis
Bibliographic record
Abstract
What Was the Question? There are conflicting calls for evidence synthesis producers to adopt or avoid recent Artificial Intelligence (AI) technologies. Our question was: How do we evaluate rapidly evolving AI tools to enhance the production of evidence syntheses and maintain quality standards? What Did We Do? To advance information retrieval science for producing evidence syntheses at Canada’s Drug Agency (CDA-AMC), the Research Information Services team developed a process to evaluate promising AI search tools. We inventoried 51 tools in the fall of 2023, established selection criteria, assessed specific attributes, and built a standalone instrument to support continuous monitoring and evaluation of new tools. What Did We Find? Rapid development of AI search tools requires a flexible evaluation instrument to inform adoption decisions and enable comparison between tools. We identified mandatory and desirable characteristics for suitable AI tools to assist with information retrieval tasks conducted by CDA-AMC. This work enabled the development of a flexible instrument to evaluate novel AI search tools for evidence synthesis. What Does This Mean? CDA-AMC operationalized a replicable process to monitor and evaluate AI search tools. Our approach to evaluating AI technologies will advance information retrieval methods by CDA-AMC, and our evaluation instrument will assist any evidence synthesis producer interested in adopting AI search 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 | 0.576 | 0.696 |
| Meta-epidemiology (narrow) | 0.002 | 0.002 |
| Meta-epidemiology (broad) | 0.005 | 0.009 |
| Bibliometrics | 0.024 | 0.020 |
| Science and technology studies | 0.003 | 0.005 |
| Scholarly communication | 0.012 | 0.013 |
| Open science | 0.003 | 0.010 |
| Research integrity | 0.004 | 0.005 |
| Insufficient payload (model declined to judge) | 0.006 | 0.002 |
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; the direct Gemma label and the distilled Codex classifier agree on what is shown here.
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".