Why it matters
RAISE states that AI in evidence synthesis should be used responsibly, transparently, ethically and with accountability. PRISMA, in turn, requires reporting methods, sources, selection processes and the use of automation tools when they affect the review workflow.
What Vetra helps document
- Tool name, module, version, date of use and affected task.
- Purpose of AI use: search, screening, extraction, critical appraisal, synthesis or drafting support.
- Which outputs were reviewed by humans and which changes were made.
- Known limitations, possible biases, assumptions and validation status.
- Data needed for PRISMA diagrams, methodological tables and transparency annexes.
How the black box is avoided
Each stage should keep a record of inputs, outputs, human decisions and changes. Traceability does not mean exposing sensitive or proprietary information; it means allowing the researcher to explain what happened and why AI use was appropriate in that context.
Final responsibility
According to RAISE, responsibility for the review remains with the researcher. Vetra reinforces that responsibility by showing suggestions, making verification easier and producing useful documentation, without attributing authorship or final judgment to the AI.