Co-Scientist AI Finds Liver Fibrosis Drug Candidate That Blocks 91% of Scarring Response
Updated
Updated · Google DeepMind · May 19
Co-Scientist AI Finds Liver Fibrosis Drug Candidate That Blocks 91% of Scarring Response
7 articles · Updated · Google DeepMind · May 19
Lab tests showed a drug candidate flagged by Co-Scientist blocked 91% of a scarring-linked response tied to liver fibrosis, pointing to a potential repurposed treatment.
Stanford's Gary Peltz used the AI system to surface overlooked drug-repurposing options, speeding a search for therapies against chronic liver disease.
The findings, published in Advanced Science, suggest gene-regulating approaches could open a new treatment path for liver fibrosis.
The result extends Co-Scientist's earlier drug-discovery work in leukemia, reinforcing its role as a multi-agent AI tool for generating testable biomedical hypotheses.
As AI generates more hypotheses, how will we safeguard against losing essential human scientific intuition?
Will AI research assistants widen the discovery gap between elite labs and smaller institutions?
If an AI makes a Nobel-worthy discovery, who should receive the ultimate credit and recognition?
Accelerating Biomedical Research: Google’s Co-Scientist AI and the Transformation of Scientific Discovery
Overview
Google's Co-Scientist is a new AI system built on Gemini 2.0, designed as a collaborative tool to accelerate scientific discovery. Developed by Google DeepMind and Edison Scientific, it addresses unmet needs in research by automating the scientific method—connecting hypothesis generation, experimental design, and data interpretation. Co-Scientist mirrors how scientists think, using advanced AI to synthesize complex information and reason through problems. Its mission is to uncover new knowledge and speed up breakthroughs, especially in challenging fields like drug discovery, by working alongside human experts and enhancing the entire research process.