Will AI Get Labor Day Off?

Labor Day weekend usually arrives with a familiar collection of traditions. There are cookouts, traffic, last trips to the beach, car sales, and the annual debate over whether summer is really over. I read Becoming Madam Secretary by Stephanie Dray (https://www.stephaniedray.com/), a historical novel about Frances Perkins, and Labor Day feels like the perfect time to revisit what her life and work still mean almost a century later.

Perkins spent her career asking what progress means if working people do not benefit from that progress.

AI is already helping us work faster, produce more, and automate tasks that once consumed hours of human effort. That raises a new Labor Day question. If AI is doing part of the work, will we eventually have to share the holiday? Will the agent expect a floating holiday, time-and-a-half, or a place at the barbecue? We should settle this before it adds itself to the community picnic sign-up sheet for potato salad.

Before Labor Day became a reason for a mattress sale

The first Labor Day celebration took place in New York City on September 5, 1882. Congress made it a federal holiday in 1894. The holiday grew from the labor movement and its fight for fairer wages, safer workplaces, reasonable hours, and recognition of workers' contribution to the country, according to the U.S. Department of Labor (https://www.dol.gov/general/laborday/history).

Nobody marched for 30 percent off patio furniture. Workers organized and negotiated for protections that we inherited without having to fight for them ourselves. Labor Day honors the people who built the country, powered its businesses, cared for its families, and insisted that a job should leave room for a life.

A productive economy should give us enough income to support our families and enough time to rest, participate in our communities, and enjoy the lives we work for.

Frances Perkins' list

Franklin D. Roosevelt appointed Frances Perkins secretary of labor in 1933, making her the first woman to serve in a presidential cabinet.

When Roosevelt asked her to join his cabinet in February 1933, Perkins arrived with terms. She gave him a nine-point list that included a 40-hour workweek, minimum wage, unemployment compensation, workers' compensation, an end to child labor, federal unemployment relief, Social Security, a stronger federal employment service, and universal health insurance, according to the Frances Perkins Center (https://francesperkinscenter.org/learn/her-life/). She literally wouldn't accept the job until the President agreed to her agenda for remaking American labor policy. Roosevelt did, and Perkins spent the next 12 years turning most of that list into American policy.

Perkins understood how power works. Laws, company policies, contracts, and negotiations determine who receives the gains from economic growth and higher output. AI changes how work gets done, but people need to manage who benefits from higher productivity.

Oh, hello, AI agent. Welcome to the team.

AI skipped the interview, ignored onboarding, and arrived in nearly every department with a suspiciously polished headshot. In July 2026, Gallup reported (https://www.gallup.com/workplace/712736/organizational-adoption-jumps-six-points.aspx) that 52 percent of U.S. workers use AI in their roles. Among employees at organizations that have implemented AI, 65 percent said it had improved their productivity and efficiency.

Marketing has become a bit of a test kitchen. AI can give me 50 headlines before I take my first sip of coffee. Fifty headlines are more than I need, but the volume highlights what AI can do. In addition to the fifty questionable headlines, I get seven cliches, six claims I need to check, and at least one sentence beginning with "In today's rapidly evolving landscape." The speed is impressive, but my judgment (and editing) still determines what gets published.

Headlines swim in the shallow end of the AI pool. In marketing, an agent can take a campaign goal, approved sources, brand rules, and access to performance data, then research competitors, organize audience findings, draft assets, build a testing plan, and monitor results. Microsoft's 2026 Work Trend Index (https://www.microsoft.com/en-us/worklab/work-trend-index/agents-human-agency-and-the-opportunity-for-every-organization) surveyed 20,000 AI users across 10 countries and found that 58 percent were producing work they could not have completed a year earlier. Advanced users were already assigning agents complex, multistep work and deciding which parts of their jobs agents should handle.

With a multistep agent, I move from editing individual deliverables to directing the entire assignment. I define the goal, choose the sources, and set the brand and business limits. The agent can execute more of the plan, and I remain responsible for the review and the outcome.

Who doesn't like a little corporate math? If AI turns a four-hour task into a one-hour task, nobody hands us the other three hours and tells us to enjoy the afternoon. By lunch, the saved time has become six new assignments and a meeting about how much capacity everyone suddenly has.

AI is fast. ROI isn't.

The time savings comes at a price. Companies pay for AI subscriptions and model usage, data cleanup, integration, security, training, and human review, often while keeping the software and processes they already had (boo, hiss). AI can cut the time required for a task while adding a new stack of costs.

PwC's 2026 Global CEO Survey (https://www.pwc.com/gx/en/1/issues/c-suite-insights/ceo-survey.html) surveyed 4,454 CEOs across 95 countries and territories. Fifty-six percent said AI had produced neither higher revenue nor lower costs during the previous year. Only 12 percent reported both.

Faster work alone proves very little about ROI. The return appears when saved time becomes lower operating costs, higher revenue, better customer retention, or work the company could not do before. Otherwise, AI is another expensive line in the software budget with much better publicity.

The pressure to justify the expense of AI lands on workers in the form of higher output targets, smaller teams, and more work. The AI vendor sends the invoice. The employee gets the new KPIs.

Who gets the hours AI saves?

Even when AI produces a return, do employees receive better pay, shorter weeks, stronger training, or more control over their work? Does faster production improve the workday, or establish a new minimum speed?

Perkins' record sets a demanding standard. Progress should leave people with better pay, safer jobs, more security, and more time. Leaders should know whether AI reduced repetitive work, improved customer service, shortened the workweek, increased profits, eliminated positions, or simply made everyone's workload, well, different.

"We became more efficient" is an incomplete answer. Efficient for whom belongs in the next sentence.

The guest list at the Labor Day cookout

AI agents can work through the holiday. Human workers have family dinners, kids parties, aging parents, picnics, and lives that require rest, dignity, and health insurance.

Labor Day remains important because work keeps changing while the basic question stays stubbornly human. Who benefits from progress?

Let the agent draft the email, summarize the meeting, monitor the dashboard, and remind us to buy hamburger buns. Give the people the holiday.

Labor Day belongs to workers. The agent can cover the inbox.