Cochrane has launched a groundbreaking study to evaluate artificial intelligence (AI) tools in evidence synthesis, aiming to determine if they can support or enhance processes like literature screening and data extraction while maintaining rigorous standards. Led by the AI Methods Group and principal investigator Gerald Gartlehner from Danube University Krems, the initiative follows a call for proposals that attracted 48 submissions, shortlisting two tools with five reserves based on alignment with RAISE principles, performance evidence, and Cochrane's values.

Employing an innovative platform design, the study will compare AI tools against traditional dual-human methods across approximately 15 Cochrane review updates. This adaptive approach allows for adding or removing tools dynamically, ensuring independent, prospective, and reproducible evaluations. Two author teams are already testing the protocol, with completion targeted for the second half of 2026. The project underscores Cochrane's commitment to responsible AI, transparency, and trustworthy evidence for better health decisions.

For the full story please click here.