Awesome Auto Research Hub Papers · Datasets · Projects
← Back to papers

A Survey of AI Scientists

arXiv 2025 59.8 method

TLDR

A survey proposing a six-stage framework for AI scientists, charting evolution from foundational modules to closed-loop systems and human-AI collaboration.

Reasoning

The paper provides a systematic synthesis of AI scientist systems with a clear six-stage framework and evolutionary timeline, which is a strength. However, as a survey, it lacks empirical evaluations or real-world experiments, limiting its direct evidence for claims.

Read-first score

Read-first score 59.8, weighted from topical fit, citation, graph, method, reproducibility, and recency signals. Original total remains 94.

Recency 8%
86.7

Uses a gentle age decay so recent papers surface without erasing older foundations. 2025

Topical relevance 42%
78.3

Uses existing LLM keyword relevance scores normalized to 0-100. AI scientist,automated scientific discovery,autonomous research agent,automated research,literature review agent,survey generation,automated experimentation,experiment design agent,AI for scientific research,paper writing agent,research automation,scientific discovery agent

Methodology quality 25%
50

Screens visible abstract and analysis fields for experiment, dataset, baseline, metric, and limitation evidence. markers=experiment

Reproducibility 25%
30

Screens links and visible text for paper, code, dataset, artifact, and repository signals. pdf=True; code=False; dataset=False; markers=none

Field roles

Frontier

Rank sensitivity

Stability: volatile; rank range: 165.

Keyword Scores

AI scientist
10
automated scientific discovery
9
AI for scientific research
9
scientific discovery agent
9
autonomous research agent
8
literature review agent
8
automated experimentation
8
paper writing agent
8
automated research
7
research automation
7
experiment design agent
6
survey generation
5

Deep Analysis

Innovations

  • Unified six-stage methodological framework deconstructing the end-to-end scientific process: Literature Review, Idea Generation, Experimental Preparation, Experimental Execution, Scientific Writing, and Paper Generation.
  • Historical charting of the field's evolution from Foundational Modules (2022-2023) to Closed-Loop Systems (2024) and Scalability, Impact, and Human-AI Collaboration (2025-present).

Methodology

The survey systematically synthesizes the AI scientist domain by applying a unified six-stage framework that decomposes the scientific workflow, then uses this lens to analyze and categorize existing systems and trends across three developmental phases.

Key Results

The survey delineates the field's progression through three distinct eras and provides a roadmap for addressing remaining challenges in robustness and governance to guide future systems.

Tags

AI