Passengers in San Francisco's autonomous taxis have encountered a disturbing new hazard: being trapped inside vehicles during violent anti-robot attacks. Doug Fulop, 37, experienced this firsthand when a man assaulted his Waymo vehicle, punching windows and threatening to kill him and his passengers for funding the technology. Unlike human drivers who could flee, Fulop's car stopped moving as designed when the attacker approached, leaving passengers helpless. This incident highlights a significant vulnerability in autonomous vehicle deployment, where safety features intended to protect can also trap victims during harassment.

Self-driving cars' programmed response to halt when people are nearby is being exploited. Videos show vandals disabling vehicles by covering sensors or spraying paint, while attackers leverage the car's immobility to threaten passengers. Fulop's six-minute ordeal ended only when bystanders distracted the attacker, allowing the Waymo to drive away. The incident raises critical questions about passenger safety protocols and the company's refusal to intervene when threatened riders call for help. Waymo maintains its policy of not manually directing vehicles away from hostile individuals, prioritizing sensor-based safety over immediate passenger rescue.

This episode threatens Waymo's public trust and growth. Fulop, a tech worker, stopped using the service after the attack, fearing nighttime rides without policy changes. Waymo's data claims a 90% reduction in serious crashes compared to human drivers, but incidents like these expose a gap between technological safety and human vulnerability. Continued expansion into 20 new cities hinges on addressing these unforeseen risks and rebuilding passenger confidence in the face of anti-robot sentiment.