Google’s AI Model Discovers New Cancer Drug Combination

On 16th October 2025, Google unveiled a family of Artificial Intelligence (AI) tools that proposed a previously unknown drug combination for detecting cancer, which proved effective in laboratory conditions.

Key Points

Introduction of C2S-Scale Model: Google introduced the Cell2Sentence-Scale 27B (C2S-Scale) — a 27-billion-parameter foundation model developed to understand the language of individual cells.
AI-Led Scientific Discovery: The model generated a novel hypothesis about cancer cell behaviour that was later experimentally validated in living cells.
Scientific Objective: The AI was tasked to detect nascent tumours that the immune system fails to recognize, by identifying a drug that boosts immune signals when interferon levels are low — a condition often seen in early-stage, undetected tumours.
Discovery of Silmitasertib: The AI identified silmitasertib, a chemical drug that enhances immune response when a tumour is suspected.
Significance of Results: The findings demonstrate that the AI’s in silico (computer-based) predictions were successfully validated in vitro (lab-based), offering a new pathway for developing combination cancer therapies.