A cluster‑randomized clinical trial published in The BMJ showed that an AI‑driven clinical decision support system improved quality of care and long‑term outcomes for acute ischemic stroke patients across participating Chinese centers. The trial reported measurable gains in adherence to evidence‑based care and functional outcomes at follow‑up, providing randomized evidence that AI CDS can affect patient trajectories in acute neurology. In parallel, a pilot deployed AI refill chatbot in Utah drew scrutiny after security researchers reported vulnerabilities in the vendor’s system, highlighting risks as health systems automate routine workflows. The incident underscores that operational AI deployments need rigorous validation, robust security reviews, and regulatory oversight to ensure patient safety while delivering efficiency gains.