AI has already delivered AlphaFold from DeepMind, predicting protein structures with unprecedented accuracy that scientists use continuously. Beyond protein folding, AI now processes image data, finds correlations in transcriptomics mapping gene expression, and analyzes massive biological datasets impossible to process manually before.
Biology today is revolutionized by AI plus CRISPR for gene manipulation and cryo-EM for viewing molecular shapes. However, Nobel Laureate Thomas Südhof warns most companies do “AI washing” — using AI to impress investors rather than achieve material advances. Biological data is abundant but not standardized, and most scientific papers contain low-quality data that turns AI output into garbage.
Südhof’s vision is clear: create AI systems that judge data quality and decide what to trust, extending transcriptomics quality parameters to all scientific data. Once biology combines transcriptomic data with CRISPR manipulation AND quality-aware AI, understanding diseases will transform. This requires integrating AI with actual biology—not just AI alone.
The next breakthrough will apply these technologies to diseases we cannot treat: Alzheimer’s epidemic and untreatable cancers like pancreatic, colon, lung, and brain cancer. Südhof believes these are now possible to approach in ways previously unfeasible because of technological innovations. Biology’s capacity to harness AI depends on standardized data, quality assessment, and maintaining America’s innovation advantage.
THE DAY AI DISCRIMINATES GOOD DATA FROM GARBLE WILL BE THE DAY BIOLOGY CURES ITS UNCURABLES.
Sanjay Sahay
Have a nice evening.