Two separate AI-enabled paths for Parkinson’s diagnosis and risk stratification surfaced. A machine learning method published in npj Parkinson’s Disease classifies individuals with Parkinson’s who are at heightened risk of falling. The study’s framing focuses directly on clinically actionable risk categorization. Meanwhile, a deep learning transcranial sonography system using cascaded super-resolution targets early-stage Parkinson’s grading, an area where detection is typically challenging before symptoms crystallize. The approach aims to improve accuracy during the window when intervention may be most valuable. Together, the reports reflect a continued shift toward imaging-and-model-driven decision support for early and safety-relevant endpoints in neurodegeneration.