Cochrane Collaboration Cautions Against AI for Science Reviews
TL;DR. The Cochrane Collaboration warned against prematurely deploying AI tools for scientific literature reviews due to accuracy and judgment concerns. - Current AI models lack the critical human judgment and expertise required for high-quality evidence synthesis. - Errors in AI-generated reviews could lead to flawed public health guidance and wasted resources on ineffective interventions. - The organization is testing AI to boost efficiency but finds existing tools unprepared for wider adoption in critical tasks.
- AI tools lack human judgment for high-quality scientific reviews.
- Errors from AI in literature reviews pose risks to public health and resource allocation.
- Cochrane Collaboration is testing AI but deems current tools unready for mainstream use.
Sources
- Why AI can't be trusted to write scientific reviews — nature.com