Only 19% of U.S. adults use AI every day, a gap that matters more in 2026 than it did even a year ago. In a televised interview, U.S. Treasury Secretary Scott Bessent offered unusually blunt career guidance for anyone entering today’s white-collar market: “I would be trained up in AI … What I would do is become an AI native.” The line lands because it frames AI as less a specialty than a new default layer of work like spreadsheets once were, then search, then cloud collaboration.
The pressure behind that advice is measurable. U.S.-based employers cited AI in 54,836 announced layoff plans in 2025, according to Challenger, Gray & Christmas. For early-career workers, the risk compounds: the unemployment rate for recent college graduates reached nearly 6% by the end of 2025’s third quarter, based on the New York Fed’s national tracking of early-career graduates. Meanwhile, the most credible macro forecasts avoid apocalyptic certainty but still point to a meaningful transition cost Goldman Sachs Research estimates 3% to 14% potential displacement under different adoption assumptions, with generative AI raising labor productivity by around 15% when fully incorporated into production.
That creates a practical question for workers: what does “AI native” look like when job titles lag behind the tools? In the simplest sense, it resembles routine fluency with general AI assistants using them to compress reading, draft first passes, reorganize messy information, and interrogate data before a human makes the final call. Consumer behavior suggests where the early leverage sits. Survey work summarized by Menlo Ventures shows writing and coding as two of the highest-penetration use cases (writing at 51%, coding at 47%), and it also reveals why many people still stall out: 63% say they do not see a need for AI in daily life, and 80% prefer interacting with people over machines. In workplaces, that preference often translates into using AI as a silent co-worker kept in the background until deadlines force experimentation.
Prompting literacy is where experimentation becomes repeatable performance. The goal is not a niche “prompt engineer” role, but a portable ability to specify constraints, iterate quickly, and keep a record of what works. Google’s Prompting Essentials course is structured around a five-step prompting framework and a reusable “prompt library” approach exactly the kind of process that converts occasional chatbot use into an operational habit. Done well, the output is not simply faster writing or cleaner slides; it is better task definition, clearer assumptions, and fewer hours spent on low-value busywork.
Bessent’s warning, stripped of TV theatrics, functions as a map for the next hiring cycle: the divide is not between technical and non-technical workers, but between those who can reliably direct AI systems and those who wait for instructions they never receive.
