A 2026 survey of 2,078 U.S. workers found that 45% have had to fix a coworker’s AI-generated work, a hidden productivity tax costing large organizations an estimated $9 million a year in rework. As AI tools spread across the workplace, the cleanup burden is reshaping how teams measure output, trust, and efficiency.
The data that got my attention
A 2026 survey of 2,078 U.S. workers by Founder Reports found that 45% have had to fix or redo a coworker’s work because it relied too heavily on AI. Among managers and senior leaders, that number jumps to 57%. The tools sold as time-savers are quietly creating a second shift of cleanup work that nobody planned for.
The Harvard Business Review, drawing on BetterUp Labs and Stanford research, put a dollar figure on it: AI-generated “workslop” costs roughly $186 per month for every affected worker. For a 10,000-person organization, that adds up to nearly $9 million a year in rework that did not exist three years ago.
Why this matters now
AI adoption has crossed a tipping point. Gallup reports that half of U.S. employees now use AI at work at least a few times a year, up from 21% in 2023. McKinsey finds that 88% of organizations use AI in at least one business function. But the rollout is outpacing the training, governance, and quality controls needed to make it work.
Only 9% of employees say they feel very comfortable using AI tools, according to Gallup. Just 25% say their employer has clearly communicated how AI should be used. Companies are deploying technology faster than they are preparing their people to use it well, and the cost is showing up in rework, mistrust, and friction between coworkers.
What the research actually shows
The data paints a more complicated picture than the productivity boost AI was supposed to deliver. Workers are not just using AI — they are scrutinizing, second-guessing, and fixing each other’s AI output. Trust in AI-assisted work is notably low.
| Metric | Percentage | Source |
|---|---|---|
| Workers who fixed a coworker’s AI-reliant work | 45% | Founder Reports 2026 |
| Managers who fixed AI-created work | 57% | Founder Reports 2026 |
| Workers who review AI work more carefully | 77% | Founder Reports 2026 |
| Workers who trust AI-assisted work less | 43% | Founder Reports 2026 |
| Workers who trust AI-assisted work more | 20% | Founder Reports 2026 |
| Desk workers encountering AI “workslop” monthly | 41% | BetterUp / Stanford / HBR |
| Rework time per AI workslop incident | ~2 hours | BetterUp / Stanford / HBR |
| Cost per affected worker per month | $186 | BetterUp / Stanford / HBR |
The scrutiny is not paranoia. The rework is real. At companies that require AI use, 73% of workers have had to fix a coworker’s AI output, and 17% say it happens regularly. Microsoft reports that 80% of the global knowledge workforce lacks the time or energy to meet current demands. AI was supposed to ease that burden. For many teams, it has added to it.
The trust gap has a demographic dimension worth noting. Workers under 40 are more skeptical of AI-assisted work than those over 50. Among younger workers, 48% report reduced trust when they learn AI was involved, compared to 34% of workers aged 50 and over. The generation most comfortable with AI tools is also the most cautious about AI output from colleagues.
A practical framework for leaders
The organizations that will see real returns from AI are the ones that treat it as a workflow change, not just a software deployment. Here is a framework leaders can use to reduce the hidden productivity tax:
- Label AI-assisted work. A simple “generated with AI, reviewed by [name]” tag on documents sets expectations and reduces surprise rework. Transparency builds trust faster than any policy document.
- Build review steps into the workflow. Do not let AI output flow directly to stakeholders without a human checkpoint. A two-minute review by the person who generated it saves twenty minutes of cleanup downstream.
- Invest in training before tools. Nearly half of employees rank training as the most important factor for successful AI adoption, yet nearly half report receiving minimal or no training. Close that gap before scaling AI further.
- Measure total workflow time, not just individual time saved. If one person saves 30 minutes using AI but two colleagues each spend 20 minutes scrutinizing the output, the team lost 10 minutes net. Track the full picture.
- Set clear quality standards. Only 25% of employees say their employer has clearly communicated how AI should be used. A one-page guideline on acceptable use, quality bars, and review expectations puts you ahead of most peers.
The bottom line
AI is not failing because the technology is weak. It is struggling because organizations are deploying it without the training, governance, and workflow design that make it productive. The 45% rework rate and the $186 monthly cost per worker are symptoms of a deeper gap between adoption and readiness. Companies that close that gap first will pull ahead. Those that do not will keep paying the hidden tax.
The World Economic Forum estimates that 59% of the global workforce will need reskilling by 2030. The organizations that start building AI literacy now — not just AI access – will be the ones whose investments actually pay off.
Where to go from here
If your organization is adopting AI but struggling with quality, trust, or rework issues, the first step is understanding where your leadership team stands on AI readiness. Most gaps start at the top. AI Leadership Readiness Assessment →
