The spotlight on autonomous vehicle safety intensified in late 2025, when multiple Waymo robotaxis were caught illegally passing stopped school buses in Austin, Texas.
Footage from Austin Independent School District revealed at least 24 such violations since the start of the school year through the middle of January, with vehicles
ignoring flashing red lights and extended stop arms while children boarded or exited. Despite a software recall in December affecting over 3,000 vehicles, incidents persisted. Investigations by both the National Highway Traffic Safety Administration (NHTSA) and the National Transportation Safety Board (NTSB) followed suit last month.
Austin ISD asked Waymo to pause operations during school hours, but the company declined, citing ongoing improvements. Director of Transportation Kris Hafezizadeh will discuss the situation next month at STN EXPO East.
This saga underscores persistent challenges in AI-driven perception systems, where even advanced neural networks struggle with contextual cues like school zones, raising alarms among educators, parents and regulators about the risks to vulnerable road users.
Power disruptions have also exposed vulnerabilities in autonomous fleets. During San Francisco’s 2025 outages, hundreds of Waymo vehicles halted abruptly, creating gridlock and highlighting dependency on stable infrastructure. Similar events in other cities have fueled debates on redundancy measures, such as onboard backup power and enhanced telematics for real-time rerouting.
As technology matures, industry experts anticipate 2026 will bring more resilient systems, with AI algorithms trained on diverse failure scenarios to minimize disruptions. Optimism persists that real-world testing will refine these tools, but incidents like these remind us that innovation must prioritize safety, especially around
schools and school buses.
The consumer automotive market is evolving rapidly, with autonomous driving features projected to become standard in over 20 percent of new vehicles this year, according to industry forecasts. Tesla’s Full Self-Driving (FSD) supervised software, for instance, has seen significant patches in 2025 and early 2026, particularly for school bus interactions. Updates have improved detection of flashing lights and stop signs, with user videos demonstrating reliable stopping and waiting behaviors.
However, NHTSA’s ongoing probes into FSD including an October evaluation of traffic law violations covering nearly 2.9 million vehicles, reveals lingering issues like occasional failures in reduced visibility. Adversarial tests by The Dawn Project staged demos showing a Tesla Model Y ignoring bus signals and striking child dummies. Tesla extended its response deadline to this month amid scrutiny of over 8,000 potential incidents. A 2023 North Carolina case, where a 17-year-old was struck by a Tesla after exiting a bus, echoes these concerns. While software fixes addressed the bug, it illustrates how AI must evolve to anticipate unpredictable child movements.
As self-driving cars proliferate in urban areas, school bus drivers face added complexity. Children in loading zones demand split-second recognition yet early AV
systems have faltered. By this year, expect wider adoption of Level 3 and 4 autonomy, where minimal human input is needed in defined conditions, promising fewer crashes
through precise sensor fusion.
NHTSA’s early 2025 estimates show overall traffic fatalities dropping: 27,365 deaths in the first nine months, a 6.4 percent decline from 2024, with the rate per 100 million vehicle miles traveled falling to 1.10. The first half of 2025 saw 17,140 fatalities, down 8.2 percent, even as miles driven rose. While distracted driving specifics for 2025 remain preliminary, trends suggest AVs could further reduce human-error crashes, though flaws in software like those in Tesla and Waymo highlight the need for rigorous validation.
Emerging trends are transforming school transportation itself. AI and telematics are shifting from reactive to proactive safety, with predictive maintenance using data
analytics to forecast bus failures, reducing breakdowns. High-definition cameras, integrated with AI software, provide 360-degree views, detecting illegal passers and
alerting authorities. Automation extends to digital forms for route planning and incident reporting, streamlining operations via cloud platforms that unify GPS/Telematics, video and RFID for student tracking.
The growth of vehicle-to-everything (V2X) communication potentially enables school buses to signal AVs directly, which could prevent illegal passes.
The school transportation industry must adapt to these innovations to safeguard students. From apps providing real-time ETA alerts to parents, to HD cameras deterring
misconduct inside buses, technology enhances efficiency and accountability.
As we share roads with evolving AVs, collaboration between manufacturers, regulators and districts is crucial. Staying ahead of the curve ensures we don’t lag in safety, after all, the families our industry serves count on us daily to innovate for the best interest of kids.
Editor’s Note: As reprinted from the February 2026 issue of School Transportation News.
Related: Investigation into Waymo Driverless Vehicles Continues Following Latest Collision with Student
Related: Waymo Driverless Vehicles Continue to Illegally Pass School Buses
Related: Waymo Driverless Car Illegally Passes Stopped School Bus in Atlanta
Related: NHTSA Investigates Autonomous Waymo Rides After Illegal School Bus Passing
