RESEARCH PROJECT · ALON GORENSHTEIN, MD
LITERAS: biomedical literature review and citation retrieval agents
LITERAS (Literature Review and Citation Retrieval Agents) is a research framework for finding biomedical publications and drafting literature reviews. Alon Gorenshtein and collaborators designed it to address an important problem in generative AI: references that are inaccurate, irrelevant, or fabricated.
Research question
Can specialized AI agents retrieve biomedical evidence and generate literature reviews with more reliable references?
Approach and evidence
The system divides the workflow among agents that plan a search, retrieve articles through the PubMed API, assess references, review the output, and synthesize a draft. Its public repository documents the architecture and setup.
The 2025 paper in Computers in Biology and Medicine evaluates citation retrieval and literature-review generation. Reported performance applies to the study's particular questions, systems, and evaluation methods; reference accuracy is not a guarantee that every claim in a generated review is supported.
Authors and contributions
Alon Gorenshtein is first author, with Kamel Shihada, Moran Sorka, Dvir Aran, and Shahar Shelly.
Research context
LITERAS is a research tool. Researchers should verify the underlying articles, citations, and synthesized claims before using its output.
Publications and resources
Research summary updated October 27, 2026.