A representative closed-loop system begins with a human-defined goal, such as finding a molecule or material with a desired property. The system's model proposes candidate designs, a robotic platform prepares and tests them, instruments measure the results, and the measurements feed back to update the model, which proposes the next round. Reported systems of this kind have produced real experimental outcomes, including improved catalysts and drug-like molecules with desired activity, rather than only computational predictions.
Can AI Run a Drug Discovery Lab on Its Own?