Two separate advances highlighted the growing use of AI in Parkinson’s diagnostics, ranging from lab-grade imaging algorithms to ultrasound-driven grading. In one report, a machine learning framework published in npj Parkinson’s Disease classifies individuals with Parkinson’s disease at higher risk of falling, offering a potential decision-support signal for risk stratification. A second report described a transcranial sonography (TCS) system powered by cascaded super-resolution deep learning designed for earlier-stage Parkinson’s grading, targeting the period when disease detection is typically least reliable. Together, the studies underscore how computational imaging and predictive modeling are converging to improve early recognition and functional risk assessment—both central to clinical workflow and trial design.