Waymo and School Buses: When AI Ignores the Law and Common Sense

Imagine an ideal world where driverless cars glide smoothly through the streets, never breaking the rules and never tiring. In that world, Waymo is the straight-A student: while Cruise crashes and exits the market, Alphabet's subsidiary methodically takes over San Francisco, Phoenix and Los Angeles. But even straight-A students have skeletons in the closet, and for Waymo they are big yellow school buses. The problem of ignoring stopping rules around school transport is not merely unsolved; it is getting worse, turning from a minor bug into a systemic threat to the future of autonomous driving.

The problem sounds deceptively simple: when a school bus stops and deploys its "Stop" sign, every vehicle around it must freeze. It is a basic safety rule, drilled into every American driver from the first lesson at driving school. Yet for Waymo's much-praised artificial intelligence, this scenario becomes an unsolvable task. Videos and reports that have become public show robotaxis confidently driving past buses with flashing lights, ignoring children at the roadside. Why does software that can recognize a flying plastic bag or a cat in the bushes fail in front of a huge yellow object?

The root of the problem lies in how neural networks perceive context. A school bus is not just a static obstacle. It is a dynamic object whose state changes within fractions of a second.

The extendable Stop sign is often splattered with mud, lit at an awkward angle or partly blocked by other cars. For a human this is not a problem: we see the context. The bus slows down, the yellow lights come on, so a stop is coming.

For Waymo's AI, which is used to relying on clear patterns, this process turns out to be too complex. If the algorithm is not 99.9% sure that it is looking at an active stop sign, it chooses to keep moving so as not to create a traffic jam out of nothing.

And that decision can prove fatal.

Regulators at NHTSA (the US National Highway Traffic Safety Administration) have already started looking closely at these incidents. After Cruise's recent troubles, any misstep by a robotaxi now draws tenfold attention. Where minor violations used to be overlooked as growing pains, the situation has changed. Ignoring a school bus is not just a ticket; it is a political and ethical trigger. No city mayor or senator wants to explain to voters why they allowed vehicles that are potentially dangerous to children to operate on the streets.

The criticism is sharpened by the fact that Waymo positions itself as the safest driver in the world. The company regularly publishes reports comparing its figures with human ones, proudly announcing a drop in crashes. But statistics are slippery. You can drive a million miles without a collision, yet a single pass by a school bus can wipe out all the credit of trust. This is a question not only of technology but of the social contract: we let robots drive among us only as long as they behave predictably and respect our most important rules.

What does this mean for the industry? Most likely, Waymo will have to urgently revise its visual recognition algorithms and possibly add specific logic layers for school zones. That will slow the service's scaling and require new investment in model training. But there is no choice: either the AI learns to respect the yellow bus, or regulators will press the "Stop" button on the entire driverless program. The irony is that a technology meant to eliminate the human factor has stumbled over a rule that any human follows by instinct.

Bottom line: If Waymo does not solve the bus problem in the coming months, a new wave of tough restrictions awaits the entire autonomous transport sector.