The Wheelchair at MIA: Automation Doesn't Delete Jobs—It Relocates Them Upstream
Automation doesn’t delete jobs. It relocates them upstream.
Terminal D, Miami International. An autonomous wheelchair rolled past me on its way from the gate to baggage claim. No pusher. A passenger sitting in a WHILL chair, following floor sensors, stopping at intersections, avoiding obstacles.
Two guys at the coffee kiosk were watching it. One said what everyone says: “See? That’s another job gone. Used to be someone pushed that wheelchair. Now it’s a robot.” Not angry. Just stating what seemed obvious.
The wheelchair didn’t appear by magic. Someone designed it. Someone wrote the navigation software. Someone built the sensors. Someone manufactured the chassis.
Someone installed the floor beacons. Someone maintains the fleet, handles support, manages the airport contract, sells the system to the next airport.
One visible job lost. Dozens of invisible jobs created upstream. The guy at the kiosk was looking at the wrong part of the picture.
The Upstream Supply Chain
WHILL has deployed at Miami International, LAX, Tokyo Haneda, and twelve other airports. Over 400,000 autonomous trips logged globally. Each deployment needed engineers, project managers, installation teams, software developers, machine learning specialists, industrial designers, regulatory specialists, sales teams, and ongoing maintenance.
The wheelchair you see is the customer-facing endpoint of an employment structure you don’t see.
Same pattern at Amazon. Everyone counts the warehouse workers whose jobs change. Nobody counts the robotics engineers in Boston, the factory workers building the robots in Massachusetts, the ML teams optimizing routes, the technicians maintaining the fleet, the analysts redesigning warehouse layouts.
Amazon runs over a million robots. It also employs 1.56 million humans. The robots created an operation large enough to require more humans, not fewer.
The Bureau of Labor Statistics projects 5.2 million new US jobs over the next decade. Computer occupations grow at 11.7%, three times faster than overall employment.
Not despite automation. Because of it. Every automated system needs a supply chain of human expertise to design, build, deploy, and improve it.
The Visibility Problem
The jobs that disappear are visible. The wheelchair pusher is someone you can see.
The jobs that appear are invisible. They sit in office parks in Silicon Valley, in factories in Michigan, in maintenance facilities you’ll never visit. You don’t see the WHILL engineer in Tokyo debugging a navigation algorithm. You don’t see the field technician replacing a worn motor.
We see the disruption. We don’t see the creation.
ATMs ran the same play. In 1985 the US had 60,000 ATMs and 485,000 bank tellers. By 2002: 352,000 ATMs and 527,000 tellers. More machines, more humans.
Cheaper banking meant more branches. Branches needed tellers for the transactions ATMs couldn’t handle.
McKinsey ran the numbers across the whole economy from 1980 to 2015. Computers eliminated about 3.5 million jobs and created 19 million. Net: +15.7 million. The jobs changed. Total employment went up.
Watching that wheelchair at MIA, I don’t see a job lost. I see an industry that didn’t exist ten years ago. Autonomous airport mobility is now a category, with companies competing in it, engineers building careers in it, standards being written.
That’s job creation. Just upstream from where most people are looking.
Why the Narrative Persists
The “automation kills jobs” story sticks because it’s emotionally compelling and locally true. If you’re the wheelchair pusher who just lost work, the macro statistics don’t pay rent.
The transition is real. The pain is real. New jobs existing somewhere else doesn’t mean they’re accessible to you.
Macro policy still can’t be driven by local pain alone. Automation relocates employment, it doesn’t eliminate it. The right response is helping people transition: retraining, safety nets, economic support. Blocking automation to preserve specific jobs freezes the economy and prevents better opportunities elsewhere.
The OECD studied task-level automation across developed economies. Only 9% of jobs are fully automatable when you decompose them into tasks.
Most jobs are bundles. AI and robots handle the automatable parts. Humans focus on judgment, creativity, and interpersonal work. The job title can stay the same; the actual work shifts.
The World Economic Forum projects a net 2% job loss through 2027 from AI and automation. Not 20%. Not 50%. Two percent. That’s a reshuffle, not a collapse.
The guy at the coffee kiosk isn’t going to run this analysis. He sees a wheelchair with no pusher and the conclusion seems obvious. For anyone watching the actual employment numbers, the conclusion is different. Automation creates jobs. Just not where you’re looking.
Common Questions
“The transition is still painful for the individuals affected.” Yes. That’s why we need better safety nets, retraining, and transition support. Net positive macro doesn’t erase micro pain. The fix is helping people adapt, not blocking the technology.
“Specific industries will still get decimated.” Travel agents, typists, telephone operators. All decimated by previous waves. Total employment still went up because new industries emerged. The question is how we manage the transition, not whether it happens.
Look upstream. The jobs didn’t disappear. They moved.