Six in ten U.S. workers who have trained employees say getting quality output from AI requires at least as much effort as onboarding a human hire, according to an October 2026 survey from Founder Reports. One in three went further, reporting that managing AI demands more work than training a person. The findings arrive as companies increasingly substitute AI tools for entry-level positions, raising questions about whether the expected cost savings actually materialize once oversight demands are factored in.
The survey polled 2,052 U.S.-based employees who use AI at least weekly, with 1,856 having direct experience training staff. Among workers who use AI multiple times daily, 56.1% said the effort level matches or exceeds human training. Those spending more than 10 hours weekly on AI oversight were especially likely to report higher demands: 39.3% said it takes more effort than training a person, compared to 27.8% of those spending under an hour weekly. Nearly one-third of respondents—31.6%—dedicate at least three hours each week to what's known as "botsitting": training, overseeing, checking, and correcting AI work. Another 8.4% spend more than 10 hours weekly on these tasks. The perception holds steady across job levels, with 62.3% of individual contributors and 58.4% of managers agreeing AI demands comparable or greater effort. Generational differences are minimal, ranging from 58.4% among Gen X to 61% among Gen Z.
Not all workers see AI as equally demanding. Forty percent of those with training experience believe AI requires less effort than onboarding humans, a proportion that climbs to 43.9% among the most frequent users. Among heavy AI users logging more than 10 hours of weekly oversight, 41% reported they would accomplish far less without these tools. Lakshmi Peri, founder of AI consulting firm Naanya, attributes the initial difficulty to fragmented organizational knowledge rather than AI limitations, noting that once systems are properly configured, the same setup can train unlimited users. Max Gladysh of BotsCrew consulting firm observes that most effort is front-loaded, with oversight eventually shifting from reviewing all output to checking only exceptions—though it will never disappear entirely. He warns that "companies get into trouble when they budget AI like a software install instead of like a new team member you keep managing."
The burden of AI oversight falls disproportionately on senior staff. Managers and above are significantly more likely to spend at least three hours weekly fixing AI errors—40.7% compared to 25% of individual contributors—and 57% have corrected colleagues' work that leaned too heavily on AI, versus 38% of non-managers. A 2026 SHRM report found directors and above spend six hours weekly correcting AI output, double the three hours individual contributors invest. As companies reduce entry-level headcount in favor of AI, this dynamic may create hidden costs. Dr. Christiane Schroeter, a Cal Poly professor and leadership strategist, argues that eliminating junior roles merely defers training expenses: "Skip that junior hire now, and you'll be recruiting a manager in five years who costs more and comes in a stranger to your business."
Quality control of AI output has emerged as the top new workplace skill, ranked first by 50% of workers in Microsoft's 2026 Work Trend Index, ahead of critical thinking at 46%. The report emphasizes that as AI expands capabilities, it simultaneously increases the value of sound judgment. Dr. Schroeter recommends that for each AI tool deployed, companies assign junior staff to examine and challenge its output, treating the exercise as critical-thinking development rather than passive acceptance. The approach addresses both immediate oversight needs and long-term talent pipeline concerns—a pipeline at risk as recent college graduate unemployment climbed to 5.7% in June 2026 from 4.2% four years prior, coinciding with prominent CEOs advocating AI-first hiring freezes. The real test for technology leaders won't be whether AI can perform entry-level work, but whether their organizations can sustain the judgment layer that makes it useful. Cutting the bottom rung may save salary costs today while creating a vacuum in the oversight roles that AI ironically makes more essential, not less.