ResearchLoop: An Evidence-Gated Control Plane for AI-Assisted Research

2026-05-27

AI Summary

While AI-assisted research accelerates the entire scientific workflow, it faces a credibility crisis due to misalignment between claims and supporting evidence. This work proposes an evidence-driven control plane that establishes a traceable and verifiable closed loop for AI-assisted research by persisting research questions, task contracts, evidence objects, and claim ledgers. The core innovation lies in the first-ever strong binding between claims and evidence, coupled with formal guarantees of reproducibility, enabled by a repository-backed runtime, declarative contracts, and a code-generation verification framework. The approach has undergone nine iterative development cycles, validated through self-hosted case studies, ablation experiments, International Mathematical Olympiad benchmarks, and SciCode boundary tests. All artifacts and verification reports are publicly released.

Abstract

AI-assisted research compresses ideation, implementation, evaluation, and manuscript writing into a single interactive loop. This compression is useful, but it also creates a publication risk: paper claims can become easier to state than to audit. We present ResearchLoop, an evidence-gated control plane for AI-assisted computational research. ResearchLoop treats research questions, task contracts, evidence objects, claim ledgers, closeouts, and paper bindings as durable project state, realized here as a repository-backed runtime. This technical report provides the complete protocol specification, state model, transition rules, claim-admission algorithm, and insight-compounding mechanism. It also reports the full experimental record spanning nine versions (V0--V9), including a self-hosting case study, a controlled task-suite study with component ablations, a mathematical olympiad evaluation, and a supplementary SciCode boundary experiment evaluated with the official generated-code harness. All artifacts, manifests, and verification reports are preserved in the project repository.

Problem

• AI-assisted research
• publication risk
• claim auditability
• evidence management
• research reproducibility

Innovation

• evidence-gated
• AI-assisted research
• claim auditing
• durable project state
• ResearchLoop