Driverless Cars, Jobless Futures?

My husband and I booked a driverless taxi called Waymo in San Francisco on our recent US trip, mostly out of curiosity. At 20:43 sharp, the car pulled up, its doors clicked open with the app. We climbed in, glanced at each other and then sat in silence as the steering wheel turned on its own and the car pulled out into the traffic.

For a few seconds, neither of us knew how to behave. Then we cracked - a burst of laughter. “It’s so weird,” Alex muttered, staring at the empty driver’s seat. The steering wheel moved, ghost-like, through a left turn. It felt as if an invisible chauffeur had been conjured out of thin air - calm, disciplined, unnervingly precise.

For a moment, we just stared, half expecting a hand to appear on the wheel. Instead, the car glided on, unbothered. The awe was instant and so was the unease of giving up control. Each time the car nosed into traffic, my body braced for a wheel-grab that never came. The surrender was unsettling at first and then, unexpectedly, relieving.

There was another odd difference. No driver meant no small talk. No chat about our plans or the weather, no polite tip at the end. Just silence. Liberating, in a way. Also a little cold.

Five minutes in, the strangeness had dulled. By our second ride, we were scrolling our phones and trusting the car to do its thing. After that, we found ourselves booking Waymo more than Uber or Lyft. What began as spectacle hardened into preference.

That speed is the tell. What begins as science fiction becomes the norm faster than we can process - but do we really understand the costs? This is the Collingridge dilemma: when a technology is new, it’s easy to shape but impossible to predict its consequences. By the time it matures and those consequences are clear, will it be too late? The roads will have been repainted, the regulations written.

In Los Angeles, the moment came not in a car but seated outside a restaurant. Whilst waiting for our meal, a small Uber Eats robot trundled into view. Its wheels negotiated the pavement, sensors blinking as it hesitated at every obstacle. A parent steered a buggy aside, two students hopped off the kerb to pass, the machine blinked dumbly, paused, then inched forward again. We leaned in, tracking every choice: how it took a pothole, how long it waited at the crossing, how close it came to a café chair. The lid finally popped with clinical precision. Efficient, yes. But there was no nod, no smile, no quick “thanks.” Just food, ferried by code.

Some places have hit pause - a regulator suspending a robotaxi permit after safety incidents, councils restricting sidewalk robots. Worker-owned delivery co-ops are experimenting with slower but fairer models. Agency exists. Sometimes it even pushes back. It’s just rarely loud.

But beneath it all sits Uber’s story: efficiency. Keep the system running, strip the resources. Elegant when the resource is fuel. Murkier when the resource is people. For companies, the calculus is brutal: machines don’t get sick, don’t take breaks, don’t ask for tips. For drivers already living with unpredictability, automation doesn’t read as progress. It reads as erasure.

Efficiency is never neutral. Even AI’s most powerful CEOs can predict faster rollouts - more sensors, better models, denser fleets. What they can’t (and often won’t) predict is what those rollouts will mean: for pay packets, high streets, public space. And those meanings aren’t technical. They’re social - decided in city halls, co-ops and regulatory rooms where the human cost of optimisation plays out in real time.

And redesign beats replacement. What might that look like? In rides: driverless on well-mapped trunk routes, with human remote assistants for edge cases and a human concierge option for vulnerable passengers. In delivery: bots handling short, level segments; humans paid fairly for stairs, safekeeping, care tasks; exceptions routed to people, not pushed to customers.

The tensions aren’t abstract. A driver refreshes his app and watches the orders vanish. Across town, a tech product manager grins as her dashboard ticks “on-time deliveries” upward. At the crossing, a mother grips her buggy tighter as a robot edges past her wheel. Meanwhile, an investor flips through a quarterly report and smiles at the rising margins. Same technology. Same city. Entirely different realities.

The future of work won’t be a clean break - all human or all machine. It will be a spectrum: some roles automated, some redesigned, some defiantly and necessarily human. And it won’t touch just one group. It impacts many at once - employees, customers, communities, regulators, shareholders - pulling them in different directions.

The companies that thrive in the age of AI won’t be the fastest to automate. They’ll be the ones that decide, in public, how to divide the work between human judgement and machine efficiency and then prove it with design, pay and accountability.

So the question isn’t whether your next burrito arrives by bike, car or robot. It’s who pays the price for that efficiency - the driver losing shifts, the city remaking its streets or all of us adjusting to a quieter, less human kind of service.

Automation is already arriving faster than regulation. Who really decides - elected officials, product teams or none of us at all?