AgentSLR: Automating Systematic Literature Reviews in Epidemiology with Agentic AI
arXiv cs.AI / 3/25/2026
💬 OpinionIdeas & Deep AnalysisTools & Practical UsageModels & Research
Key Points
- The paper presents AgentSLR, an open-source agentic AI pipeline that uses large language models to automate systematic literature reviews in epidemiology from retrieval through screening, data extraction, and report synthesis.
- In experiments on epidemiological reviews for nine WHO-designated priority pathogens, AgentSLR reportedly matches expert-curated ground truth performance while cutting end-to-end review time from about 7 weeks to around 20 hours (~58× speed-up).
- A benchmark across five frontier models suggests that SLR performance depends more on each model’s distinctive capabilities than on model size or inference cost alone.
- The authors include human-in-the-loop validation to identify key failure modes, highlighting where agentic automation may still need supervision.
- Overall, the study argues that agentic AI can substantially accelerate specialized scientific evidence synthesis, potentially reducing bottlenecks for evidence-based policy.
Related Articles
Build a WhatsApp AI Assistant Using Laravel, Twilio and OpenAI
Dev.to
Santa Augmentcode Intent Ep.6
Dev.to
Your Agent Hired Another Agent. The Output Was Garbage. The Money's Gone.
Dev.to
Anthropic shut down the Claude OAuth workaround. Here's the cheapest alternative in 2026.
Dev.to
ClawRouter vs TeamoRouter: one requires a crypto wallet, one doesn't
Dev.to
