Model Hacking ADAS to Pave Safer Roads for Autonomous Vehicles

**Research summary:** McAfee ATR demonstrates adversarial-machine-learning (model hacking) attacks against MobilEye traffic-sign recognition used in older Tesla vehicles, showing physical sticker perturbations can cause misclassification of speed-limit signs and — in vehicles using Speed Assist/TACC — can trigger unintended autonomous acceleration; findings were responsibly disclosed and newer MobilEye hardware appears to mitigate the issue.