What the pandemic taught me about automation

Eight months ago, the automation conversation had a comfortable rhythm.

Should we automate? What about jobs? What about the human element? These were conference panel questions. They came with wine and a moderator and polite disagreement. The timeline was vague. “Someday.” “In the coming decades.” “When the technology matures.”

The technology matured in March.

The list

I’ve been keeping a running list of automation milestones that happened because of COVID. Not things that were already planned. Things that accelerated because a virus made human presence dangerous or impossible.

Amazon added thousands of warehouse robots. Not as a future plan. As an immediate response to keep fulfillment running while human workers called in sick or stayed home.

Nuro got emergency authorization from the FDA to use its autonomous vehicles for medical supply delivery. A company that was doing pilot programs became an essential service overnight.

Hospitals deployed Intuitive Surgical robotic systems for procedures that could reduce the time patients spent in operating rooms, reducing exposure risk for medical staff.

Restaurants installed ordering kiosks. Grocery stores deployed self-checkout. Factories accelerated plans for reduced-headcount production lines. Airports put in temperature-scanning robots.

None of this is individually surprising. All of it together, in eight months, is staggering.

The decade that happened in a year

McKinsey published a report suggesting that the pandemic pulled forward a decade of automation adoption. I think they’re right, and I think the mechanism is simple.

Before COVID, automation was a cost optimization. You automated to save money, and the savings had to justify the investment over 3-5 years. The decision was rational, spreadsheet-driven, slow.

After COVID, automation was a survival requirement. You automated because humans couldn’t be in the room. The justification wasn’t ROI. It was continuity. The decision was urgent, emotional, fast.

Urgency changes everything. A technology that would have been adopted in 2028 got adopted in 2020 because someone in a supply chain operations center said “we can’t get enough people in the warehouse, what else can we do?”

What doesn’t go back

This is the part I keep thinking about.

When the pandemic ends, some of this reverses. People go back to offices. Restaurants remove the plexiglass. We shake hands again. The things that were temporary become temporary.

But the robots don’t get returned. The automated warehouse doesn’t get un-automated. The self-checkout kiosks don’t get replaced with human cashiers. The delivery robots don’t go back in the lab.

Why would they? They work. They showed up every day during a pandemic. They didn’t get sick. They didn’t need PPE. They didn’t call in because their childcare fell through.

I’m not saying this as a celebration. I’m saying it as an observation. The automation that got deployed in crisis mode will stay deployed in normal mode. Because once you’ve proven a machine can do the job, the economic case for keeping it never goes away.

The question we skipped

We never finished the debate.

Should we automate? What about the workers? What about the social contract? Those questions had complicated, careful answers that we were slowly, carefully working through.

The pandemic said “never mind, we’re doing it now” and the careful working-through got replaced by emergency deployment.

The workers still need answers. The social contract still needs updating. But now we’re having those conversations after the fact, with the robots already in place, rather than before.

I don’t know if that makes the answers harder or easier. Both, probably.

I just think we should be honest about what happened. We didn’t choose to accelerate automation. A virus chose for us. And now we’re living with the consequences of a decision we didn’t make, which is a very strange position for a civilization that prides itself on agency.

Related thinking: