When David Osei was laid off from his data entry position at a logistics company in Atlanta last January, his manager handed him a folder of severance paperwork and a brochure for a retraining program. The cause of the layoff wasn’t a downturn. The company had deployed a new AI automation system that could handle David’s entire workload — and the workload of 14 of his colleagues — in a fraction of the time.
“The weird part,” David said, “is I’d actually helped them set it up.”
Stories like David’s are rippling through the American workforce. A 2025 McKinsey report estimated that nearly 12 million U.S. jobs could be disrupted by AI automation over the next three years, with clerical, administrative, and entry-level tech roles bearing the heaviest impact. Customer service, data processing, paralegal work, and basic coding tasks are already seeing significant automation-driven shifts.
But the picture is genuinely complicated. For every job eliminated, economists point to new roles emerging — prompt engineers, AI trainers, human-AI liaison specialists, and automation auditors are among the fastest-growing positions on LinkedIn. The challenge isn’t that work is disappearing entirely. It’s that the new work requires different skills, and the transition is happening faster than training pipelines can keep up.
The federal government has responded with a $4 billion “AI Workforce Transition Initiative,” funding community college programs and apprenticeships focused on tech-adjacent skills. Critics argue it’s too little, too late. Supporters say it’s a start.
David, for his part, enrolled in a six-month data analytics bootcamp and recently landed a contract role reviewing AI-generated logistics reports for accuracy. His income is about 15% lower than before, but he’s optimistic.
“The machine does the first draft,” he said. “I tell it where it’s wrong. That feels like a job worth having.”
