Scholar Rebuts 'AI Will Cure Cancer': Better Molecules Can't Fix Wrong Targets

anshulkundaje · x · 2026-08-05

A pioneer in AI-in-biology has published a sharp rebuttal to the "AI will cure cancer" narrative. The core argument states that the current bottleneck in drug discovery is fundamentally a measurement problem.

High clinical failure rate: Over 90% of drugs entering clinical trials still fail, a number unchanged for decades. In most cases, the molecule works fine, but the mechanism it was designed to target was simply wrong.
Limitations of AI: Designing better molecules (better keys) via AI won't open the correct doors (targets) if the underlying mechanism is flawed.
Missing data: Human biology has never been measured at the scale or resolution needed to actually map disease mechanisms. LLMs can connect existing literature, but they cannot extract answers about molecules and processes that were never captured in the first place.
The role of proteomics: Scalable mass spectrometry-based proteomics can directly quantify these missing elements, which is where the true solution lies.

Original post →

Scholar Rebuts 'AI Will Cure Cancer': Better Molecules Can't Fix Wrong Targets

More from Research