AI accelerates iteration and improves selection, but humans still validate.
AI gives us better priors, better search, better filters, and faster loops. It narrows the hypothesis space, surfaces weak signals earlier, improves decision support, and accelerates the cycle of data to model to candidate to experiment and back to data. The headline-grabbing examples from Exscientia and Insilico Medicine don’t mean AI replaces scientists. They fit this pattern: AI helps prioritize, propose, and optimize, while humans validate in wet labs and the clinic. The real impact shows up primarily as faster iteration and better selection under uncertainty, not autonomous discovery. It makes teams materially more effective.
— Healthcare’s AI Lesson: Autocomplete Isn’t Understanding · Forbes