A Widespread Trust Deficit
The core of the issue isn't whether employees are using AI, but how they trust its output. The headline figure, suggesting two-thirds of employees double-check AI's work, reflects a broader trend seen across multiple recent studies. For instance, a May
2026 report from Founder Reports found that 77% of workers review a coworker's work more carefully when they know AI was used. Another survey from August 2026 by Hint App noted that 64% of employees spend additional time reviewing AI-generated material. This isn't a case of a few skeptical individuals; it’s a systemic behavior. Employees are embracing AI as a tool but are unwilling to accept its output on faith, creating a new, hidden layer of work focused entirely on verification.
The Specter of 'AI Slop'
The primary driver behind this verification habit is the inconsistent quality of AI-generated content, often termed 'AI slop'. Employees are wary of inaccuracies, subtle biases, and a general lack of contextual nuance that can make AI output unreliable or even damaging. Research from July 2026 found that 46% of workers feel that dealing with low-quality AI output makes their job less meaningful. More telling, 45% of workers admit they've had to fix or completely redo a coworker's task that relied too heavily on AI. This experience of cleaning up after the machine has fostered a 'trust but verify' mentality, which in many cases has morphed into 'distrust and meticulously verify'.
The Hidden Cost of Verification
The promise of AI has always been rooted in productivity gains and time savings. However, this widespread need for verification presents a productivity paradox. If an employee saves an hour using an AI tool but then spends 45 minutes fact-checking and refining the output, the net gain is minimal. One study highlighted this exact issue, finding that 42% of workers now spend more time verifying AI output than they save by using it. This extra review time is a hidden cost that many organizations fail to account for when measuring the return on investment for AI tools. It shifts the burden of work from creation to validation, and as one report notes, if multiple colleagues are scrutinizing AI-assisted work, a team can easily end up with a net loss in time.
From Blind Trust to Critical Partnership
While double-checking is born of distrust, it also points toward a more mature, sustainable model for human-AI collaboration. Instead of viewing AI as an infallible oracle, savvy employees are treating it as a highly capable but error-prone assistant. This approach, often called the 'human-in-the-loop' model, leverages AI for first drafts, data processing, and idea generation, while reserving human expertise for refinement, critical thinking, and final judgment. A 2026 Epoch AI survey found that two-thirds of AI-generated outputs are used with only minor edits or as-is, suggesting workers are finding a balance. The act of verification, in this light, is not just a tedious chore but a critical part of the new workflow—a quality control step that ensures the final product meets human standards.
What Leaders and Managers Must Understand
This trend has clear implications for leadership. Simply deploying AI tools and expecting productivity to soar is proving to be a flawed strategy. Managers need to acknowledge the verification burden and provide clear guidance. A stunning 44% of U.S. workers report their employer has no clear AI policy. Without guardrails, employees are left to navigate this new terrain alone. Leaders must invest in training that goes beyond how to use AI and focuses on how to critically evaluate and ethically manage its output. Fostering a culture of transparent AI use—where employees are encouraged to disclose when and how they use the tools—can also help rebuild trust and turn individual verification into a collaborative process of quality assurance.
