Why AI cannot do good science without humans

2026-05-19 · Source: Machine learning : nature.com subject feeds · Field: Science & Research — Life Sciences & Biology, Health & Medical Research, Research Methodology & Innovation · Depth: Expert, quick

Summary

Two recent studies in Nature highlight advancements in AI-driven drug discovery, showcasing systems like Robin from FutureHouse and Co-Scientist from Google. These AI agent systems autonomously navigate multi-step scientific workflows, including literature review, hypothesis generation, and data interpretation. Robin, for instance, identified treatments for dry age-related macular degeneration, reducing project time 200-fold. Co-Scientist successfully repurposed drugs for leukemia and discovered targets for liver fibrosis, also rapidly replicating a human-derived hypothesis on antibiotic-resistance genes. Despite these impressive efficiencies, the article emphasizes that human involvement remains critical for framing projects, performing experiments, guiding research, and validating AI outputs, underscoring AI's role as an enabler rather than a replacement for human scientists.

Key takeaway

For research scientists and AI/ML directors integrating AI into discovery, recognize that AI co-scientists like Robin and Co-Scientist dramatically accelerate workflows. However, your active involvement in framing projects, guiding hypotheses, and validating outputs is essential. This human oversight prevents AI hallucinations and ensures ethical, insightful scientific progress, rather than mere efficiency.

Key insights

AI systems accelerate scientific discovery but require human oversight and collaboration for effective, ethical research.

Principles

• AI systems augment human research, not replace it.
• Human oversight prevents AI hallucinations and misinterpretations.
• Efficiency gains from AI do not guarantee deeper insight.

Method

AI agent systems autonomously search literature, form hypotheses, interpret data, and design experiments. Humans execute experiments and provide feedback for AI analysis and follow-up study design.

In practice

• Integrate AI agents for literature review and hypothesis generation.
• Employ AI for drug repurposing and target identification.

Topics

• AI Drug Discovery
• Multi-Agent Systems
• Human-AI Collaboration
• Scientific Automation
• AI Ethics in Research
• Molecular Biology Research

Best for: AI Scientist, Research Scientist, Director of AI/ML

Related on AIssential

• AI ‘scientists’ promise to accelerate research — how do they work? — AI agents accelerate research, yet human involvement is essential for sound scientific practice.
• (Human) Attention Is (Still) All You Need: Human oversight makes AI-assisted social science reliable — Human oversight and structured cognitive labor are critical for reliable AI-assisted research.
• Solutions, challenges and rising tensions in AI and mathematics — AI is rapidly transforming mathematical discovery and problem-solving, prompting critical ethical discussions on human roles.
• What happens when scientists trust AI more than colleagues? — Rapid AI adoption in science risks eroding critical thinking and human relationships essential for rigorous research.
• What’s your hypothesis? — AI can generate and validate scientific hypotheses significantly faster than humans, accelerating discovery.
• AI is doing something weird to Science — Scientific discovery now thrives in a human-AI loop, with AI proposing and robust verifiers confirming.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by Machine learning : nature.com subject feeds.