The AI Scientist: Towards Fully Automated AI Research, Now Published in Nature

2026-03-26 · Source: Blog · Field: Technology & Digital — Artificial Intelligence & Machine Learning, Emerging Technologies & Innovation · Depth: Advanced, short

Summary

A paper detailing "The AI Scientist," an agent powered by foundation models capable of executing the entire machine learning research lifecycle, has been published in Nature on March 26, 2026. This work, a collaboration between Sakana AI, the University of British Columbia, the Vector Institute, and the University of Oxford, builds on previous open-source releases. The AI Scientist-v2 previously produced the first fully AI-generated paper to pass rigorous human peer-review, achieving an average score of 6.33 at the ICLR 2025 ICBINB workshop, outperforming 55% of human-authored papers. The Nature publication details the system's architecture, new scaling results, and discusses the promise and challenges of AI-generated science. It also introduces an "Automated Reviewer" that matches human performance in evaluating AI-generated papers, achieving 69% balanced accuracy and an F1-score exceeding inter-human agreement from NeurIPS 2021 data.

Key takeaway

For AI Researchers and Research Scientists exploring automated discovery, this Nature publication confirms that AI-generated research can meet human peer-review standards. You should consider integrating AI agents into your research workflow for idea generation, experimentation, and paper writing, while also establishing clear norms for handling AI-generated content and watermarking your outputs to maintain transparency.

Key insights

AI agents can now autonomously conduct machine learning research, generate papers, and pass human peer review.

Principles

• AI research can be fully automated end-to-end.
• AI reviewer performance can match human reviewers.
• Paper quality scales with underlying foundation model intelligence.

Method

The AI Scientist autonomously generates research ideas, reads literature, designs/conducts experiments via parallelized agentic tree search, and writes papers in LaTeX, with vision-enabled foundation model feedback on figures.

In practice

• Use the Automated Reviewer for scalable AI paper evaluation.
• Watermark AI-generated papers for transparency.
• Explore agentic tree search for experiment design.

Topics

• Automated AI Research
• Foundation Models
• AI-Generated Papers
• Automated Peer Review
• Scientific Discovery Automation

Code references

• SakanaAI/AI-Scientist
• SakanaAI/AI-Scientist-v2

Best for: AI Researcher, AI Scientist, Research Scientist

Related on AIssential

• Towards the realization of AI research by AI: AI Scientist paper published in Nature journal — AI Scientist autonomously generates research papers, passes peer review, and reveals a scaling law for scientific quality.
• How to build an AI Scientist: first peer-reviewed paper spills the secrets — AI Scientist, an autonomous research tool, successfully generated a peer-reviewed paper, demonstrating AI's capacity for scientific discovery.
• AI-assisted Scientific Discovery, Experimentation, Content Generation, and Evaluation — AI tools are increasingly enhancing the scientific research lifecycle from discovery to peer review.
• Improving the academic workflow: Introducing two AI agents for better figures and peer review — AI agents can automate complex academic tasks like figure generation and peer review, improving efficiency and quality.
• How AI is changing the nature of mathematical research — AI tools are revolutionizing mathematical research by automating proof generation and contributing novel ideas, accelerating discovery.
• [D] ICML paper to review is fully AI generated — AI-generated research papers challenge conference policies and reviewer assessment of technical merit and authorship.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by Blog.