Two AI-based science assistants succeed with drug-retargeting tasks

2026-05-19 · Source: AI - Ars Technica · Field: Science & Research — Life Sciences & Biology, Health & Medical Research, Research Methodology & Innovation · Depth: Intermediate, medium

Summary

Nature recently published two papers detailing new AI systems, Google’s Co-Scientist and FutureHouse’s Robin, designed to aid scientists in developing and testing hypotheses, particularly for drug retargeting. Co-Scientist, based on Google’s Gemini LLM, operates as a "scientist in the loop" system, evaluating hypotheses for plausibility, novelty, testability, and safety through a "tournament" and iterative refinement agents. FutureHouse’s Robin, an agentic system, utilizes specialized tools like Crow for summaries and Falcon for deep overviews, enabling it to analyze 551 papers in 30 minutes, a task estimated to take a human 540 hours. Robin distinguishes itself with its Finch tool, which automates data evaluation from standard biological screening assays such as flow cytometry and RNA-seq. Both systems successfully suggested known drugs for conditions like acute myeloid leukemia and macular degeneration, demonstrating their capability to identify non-obvious connections across disparate scientific literature.

Key takeaway

For Research Scientists grappling with literature overload in drug discovery, these AI assistants offer a critical advantage. You should consider integrating such agentic systems to identify non-obvious connections and generate testable hypotheses, especially for drug repurposing. This approach can significantly reduce the time spent on literature review and accelerate the initial stages of drug development, allowing you to focus on complex experimental design and validation.

Key insights

AI-powered agentic systems can accelerate scientific discovery by synthesizing vast literature and generating testable hypotheses.

Principles

• "Scientist in the loop" models enhance AI reliability.
• Agentic AI excels at combinatorial synthesis of knowledge.
• Specialized literature tools reduce AI hallucination.

Method

Both systems interpret research goals, conduct literature searches, form hypotheses, and evaluate them through iterative processes like "tournaments" or pairwise comparisons, often involving human expert review. Robin adds automated assay data evaluation.

In practice

• Use AI to identify drug repurposing candidates.
• Apply agentic systems for cross-disciplinary literature review.
• Integrate automated data evaluation for biological assays.

Topics

• AI in Drug Discovery
• Agentic AI Systems
• Hypothesis Generation
• Literature Review Automation
• Drug Repurposing
• Google Co-Scientist
• FutureHouse Robin

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

Related on AIssential

• Teams of AI agents boost speed of research — AI agent teams can rapidly generate plausible scientific hypotheses and identify drug candidates, significantly accelerating early-stage research.
• 3 Questions: Using AI to accelerate the discovery and design of therapeutic drugs — AI, particularly deep learning and generative models, significantly accelerates novel antibiotic discovery and design.
• AI is reengineering drug discovery by speeding up testing and scanning petabytes of data for connections between diseases — AI significantly enhances drug discovery by analyzing vast data to predict targets, efficacy, and disease interrelationships, improving success rates.
• He Raised $70M to Cure Every Disease With AI — AI agents, particularly specialized models, significantly accelerate scientific discovery by removing human talent bottlenecks and enabling high-throughput reasoning.
• Accelerating discovery of liver disease mechanisms — AI-driven literature analysis can uncover novel biological hypotheses and candidate therapies for complex diseases.
• TAI #192: AI Enters the Scientific Discovery Loop — AI models are transitioning from tools to active participants in scientific discovery, particularly when coupled with robust verification.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by AI - Ars Technica.