Generative AI tools are rapidly transforming how software is built—and raising new risks in the process, according to a new TechBrief from the Association for Computing Machinery’s Technology Policy Council (TPC) on the rise of “vibe coding.”

A research team has developed a camera-only visual odometry (VO) system that uses prebuilt colored point cloud maps to deliver more accurate and robust localization with reduced drift in Global Navigation Satellite System (GNSS)-challenged environments. 

Researchers at Concordia University have developed an AI-driven robotic system that can autonomously perform cardiac ultrasound scans without human guidance. Trained using realistic simulated images and reinforcement learning, the robot learns to position and adjust the probe to capture diagnostic views. Tests show it can match or outperform human operators, potentially expanding access to reliable heart imaging in underserved settings. 

UC San Diego researchers have developed T1GRS, a machine‑learning model that accurately predicts genetic risk for Type 1 diabetes, paving the way for earlier diagnosis and more personalized treatment.