AI Accuracy

What if the AI misses a relevant paper?

Use the Generous philosophy for abstract screening. It biases the AI toward inclusion, so you're less likely to miss anything. You can apply stricter criteria later at the full-text stage.

Can I verify the AI's accuracy before screening?

Yes. The calibration run lets you screen 20 papers manually, then the AI screens the same 20 blindly. You can compare the results before committing to a full screen.

Does the AI improve over time?

Your overrides are stored and can inform future screening runs. The more you use Oryn, the better it understands your review criteria.

What if I disagree with a decision?

Click the paper, review the AI's reasoning, and click a different decision. Your override is final and logged in the audit trail.