The Hidden Economics of Food - Part 16: The Robot in the Field

The Robot’s Limit

Strawberries are delicate. They bruise easily. They ripen at different rates on the same plant. They hide under leaves.

Despite decades of research, strawberries are still picked by hand. Human workers—often migrant workers, often underpaid—bend over fields, identify ripe berries, and pick them without crushing.

This should be reassuring. Technology can’t do everything. Some human skills remain irreplaceable.

But it’s also concerning. If technology did replace strawberry pickers, what would those workers do? And what happens when the robots improve?

The Automation Story

Technology has always changed work:

Agriculture

In 1800, 80% of Americans worked in agriculture. Today, it’s under 2%. Tractors, harvesters, and chemicals replaced human labor. Food became cheaper. Displaced farmers moved to cities.

Manufacturing

In the mid-20th century, manufacturing employed huge workforces. Automation—assembly lines, robots, computer control—eliminated many of those jobs. Production increased; employment fell.

Services

Now automation is reaching services:

• Self-checkout replaces cashiers

• Online banking replaces tellers

• Automated customer service replaces call centers

• AI writing tools may affect knowledge work

Each wave eliminates jobs that seemed irreplaceable until technology replaced them.

The Luddite Question

When the Luddites smashed textile machines in early 19th-century England, they were dismissed as opponents of progress.

But were they wrong?

In the Short Run: They Had a Point

Textile automation did destroy livelihoods. Skilled weavers became unemployed. Families suffered. Communities collapsed.

The claim that “the market will adjust” was cold comfort to those who didn’t survive to see the adjustment.

In the Long Run: New Jobs Emerged

Eventually, the economy created new jobs—in factories, in services, in industries that didn’t exist before.

But this took decades. And the new jobs weren’t necessarily better than the old ones. Factory work was often brutal, dangerous, and poorly paid.

The Pattern

Technology destroys existing jobs, then (sometimes) creates new ones. The transition is painful. And whether the new jobs are good depends on labor power, not just technology.

Why Strawberries Resist

Strawberry picking illustrates what’s hard to automate:

Physical Dexterity

Human hands are remarkably precise. Robots that can match human dexterity exist but are extremely expensive.

Judgment in Variable Conditions

Strawberries vary. Each plant is slightly different. Ripeness depends on color, feel, and context. Humans handle this naturally; machines struggle.

Unstructured Environments

Factories can be designed for robots—standard layouts, predictable conditions. Fields are messier. Weather, terrain, and plant growth create variability.

The Cost Comparison

Human strawberry pickers are paid poorly. This is terrible for workers but removes incentive for automation. Why develop expensive robots when cheap labor is available?

When Automation Wins

Other agricultural tasks have been automated:

Grain Harvesting

Combines harvest wheat, corn, and rice efficiently. The crops are uniform, the fields are flat, and the work is suited to machines.

Processing

Once harvested, food is processed in factories—canning, packaging, freezing—where automation is straightforward.

Dairy

Robotic milking systems let cows choose when to be milked. This works because the task is standardized and cows are cooperative.

The pattern: automation succeeds when tasks are standardized, environments are controlled, and precision requirements are moderate.

What AI Changes

Previous automation threatened manual labor. AI threatens cognitive work:

Routine Cognitive Tasks

Tasks that follow rules—data entry, basic analysis, standard reporting—are increasingly automated.

Pattern Recognition

AI can identify patterns in data—diagnose diseases from images, predict equipment failures, detect fraud. This affects doctors, engineers, analysts.

Content Creation

AI can generate text, images, code, and music. Not perfectly—but good enough for many purposes. This threatens writers, designers, programmers.

Decision Support

AI can recommend decisions—what to buy, whom to hire, what to produce. Humans still decide, but AI shapes the decision.

The new automation doesn’t just affect manual workers. It affects professionals and knowledge workers too.

The Distribution Question

Whether automation helps or harms depends on who captures the gains:

Scenario A: Workers Share Gains

If productivity gains go to workers—through wages, shorter hours, or better conditions—automation is a blessing:

• Same output with less work

• Higher incomes

• More leisure

• Reduced drudgery

This happened in the mid-20th century, when strong unions and full employment policies ensured workers benefited from productivity growth.

Scenario B: Capital Takes All

If productivity gains go only to capital—through profits, executive pay, or shareholder returns—automation creates:

• Unemployment or precarity

• Stagnant wages

• Overwork for some, no work for others

• Social instability

This is closer to recent experience. Productivity has risen; median wages have stagnated; returns have gone to shareholders and executives.

The Difference Is Policy

Technology doesn’t determine outcomes. Policy does:

• Taxation can redistribute automation gains

• Labor law can protect worker bargaining power

• Education can enable transitions

• Social insurance can protect the displaced

• Working time regulations can spread work

The question isn’t whether to automate—that’s often not a choice. The question is how to share the gains.

The Strawberry Workers

Who picks your strawberries?

Often: migrants, undocumented workers, people with few other options. They work long hours, in difficult conditions, for low pay.

If automation replaced them, what would happen?

Without Good Policy

Workers would lose livelihoods with no alternative. They’d compete for other low-wage jobs, driving wages down. Poverty and precarity would increase.

With Good Policy

Workers might transition to other work—supported by retraining, income support, and new job creation. Or they might share in automation’s gains through higher wages or shorter hours while machines were still limited.

The technology doesn’t determine which happens. Policy does.

The Future of Work

Several possibilities exist:

Optimistic

AI and automation boost productivity. The gains are shared. People work less and live better. New activities—creative, caring, community-building—become more central.

Pessimistic

Automation eliminates jobs faster than new ones appear. Gains go to capital owners. Mass unemployment or precarity spreads. Social stability is threatened.

Mixed

Some workers benefit; others suffer. Inequality widens. Political conflict intensifies. The future is negotiated through struggle.

The optimistic scenario is possible—but not automatic. It requires deliberate policy. Without it, the pessimistic or mixed scenarios are more likely.

What Strawberries Teach

Strawberries remind us that:

• Automation isn’t all-or-nothing—it proceeds unevenly

• Human capabilities remain valuable, but that can change

• Low wages can prevent automation (by making machines uneconomical)

• Technology doesn’t determine social outcomes—policy does

The delicate berry that robots can’t pick is a pause in an ongoing process. The respite won’t last forever.

What matters is what we do before the robots learn—building systems that share automation’s gains rather than concentrate them.

The strawberry pickers deserve a future. Whether they get one depends on politics, not technology.