A screenshot posted on Reddit earlier this month captured a curious DoorDash order: $6.25 to travel about a mile and close the door of a Waymo self‑driving vehicle, with an extra $5 paid after verification. Waymo and DoorDash later confirmed the order was real and part of a small Atlanta pilot in which nearby delivery couriers are alerted when an unattended vehicle cannot resume service because a passenger failed to close a door properly.
The technical issue is simple but operationally painful. Many autonomous vehicles will not begin driving if a door sensor registers that a door has not latched securely, leaving the car stranded, out of service and potentially blocking traffic while engineers are dispatched. Paying a local courier to pop by and close a door is a low‑latency, low‑cost workaround that gets the vehicle back on the road quickly.
This episode underscores a broader truth about so‑called “driverless” systems: the most advanced autonomy still relies heavily on humans for edge cases, physical maintenance and quick fixes. Waymo, Alphabet’s autonomous driving arm, has used similar arrangements before — late last year it contracted with Honk in Los Angeles to have attendants close doors for fees reported as high as $24 — and the company says future vehicle designs will include automatic door‑closing, without giving a precise timetable.
There is a clear commercial logic. Waymo completed a large financing round this month — reportedly about $16 billion, valuing the company at roughly $126 billion — and a stalled car is lost revenue, a traffic hazard and a reputational risk in the competitive robo‑taxi market. For a well‑funded operator, modest payments to gig‑economy workers are a cheap way to protect uptime and customer experience while engineering solutions are rolled out.
But the workaround raises policy, safety and labor questions. Outsourcing ad hoc vehicle maintenance to couriers creates new micro‑tasks that sit somewhere between delivery and light field service; companies must consider liability if someone is injured while attempting to close a door, as well as background checks, training and data‑protection issues when drivers enter, inspect or touch a passenger vehicle.
The practice also highlights how the growth of automation is spawning hybrid work arrangements rather than pure substitution. Firms from warehouses to food delivery have found humans indispensable for exceptions management; in cities where curb space and traffic flow are tightly managed, the social cost of a single disabled vehicle can be disproportionate. Regulators and city planners will need to grapple with these mixed human‑robot systems as autonomous fleets scale up.
For now the arrangement is tactical rather than strategic: Waymo prefers engineered fixes like automated door actuators, but until such hardware and software are universally deployed, the gig economy provides a flexible, market‑priced buffer. How operators price these interventions, document them for safety and scale them without creating precarious, unsafe gigs will be an early test of whether the autonomous‑vehicle era reduces or simply reshuffles on‑the‑ground human labor.
