Curing disease with AI: why science is the most agreed-on upside
September 27, 2026 · 18 min read
The Chan Zuckerberg Initiative, Dario Amodei and Lee Cronin all see AI speeding up discovery in biology and chemistry, from cell atlases to automated chemistry experiments. Even cautious leaders point to science as the clearest win, which makes it the strongest case for continuing development.
The rules, numbers and steps worth keeping in front of you.
- Shared view: faster science is the AI upside that survives the disagreements.
- Chan and Zuckerberg: most big discoveries followed a new tool or a new way of seeing.
- CZI's goal: help science cure, prevent or manage all disease by the end of the century.
- Chan: single-cell data is too large for people to read, but large AI models can.
- Zuckerberg: use a model's imagined possibilities as an advantage, then have scientists check them.
- Chan: mice are not humans, so results from computer experiments must translate.
- Amodei hopes to compress progress expected by 2100 into roughly 2027 to 2032.
- Amodei says that timeline is not a scientific prediction.
- Amodei: AI cannot skip physical limits or human institutions.
- Cronin calls AI doom nonsense, yet his lab's model designs molecules for a robot to make.
- Cronin: today's models interpolate from the past and don't yet produce true novelty.
- Sejnowski: doctors and AI each scored about 90 percent; together they reached 98.
- Sejnowski: AI covers breadth and rare cases, the expert covers depth.
- For health decisions, AI tools are no substitute for a qualified clinician.
This is the short form. The full insight has every expert's reasoning, their own words and the sources.