Review
Open access
Apr 2026
Comparison of three large language models' ability to assess the risk of bias using ROBINS-I tool
Current off-the-shelf LLMs cannot reliably perform ROBINS-I risk of bias assessments, as Claude was internally consistent but poorly aligned with human reviewers and GPT had lower reliability and mixed accuracy.
S. Saadi, Farah Fleti, Omar H. Rajjoub et al.
· BMJ digital health & AI · 0 citations