Our Take
Half your workforce thinks it's expert at AI while one in four fears admitting they use it at all—the real risk isn't adoption speed, it's the silence around it.
Why it matters
Organizations are deploying AI policy in a vacuum. Employees are adopting tools without guardrails, losing the analysis and collaboration that made the work valuable in the first place. Leadership silence creates fear, not safety.
Do this week
CHRO: Publish explicit AI use guidelines and have your leadership visibly document one AI tool they use this week so employees see transparency modeled before they're asked to adopt it.
The adoption-guidance gap
A survey by INTOO (conducted by The Harris Poll) found that 52% of employees consider themselves AI experts for work tasks, and 63% believe their AI knowledge makes them more valuable (per company-reported data). Yet clarity is absent: 20% say they don't know what's acceptable when using AI at work, and 25% report discomfort disclosing AI use to colleagues.
The disconnect deepens when workers face social friction. Forty-two percent say it would be embarrassing to ask colleagues for help with new technology, including AI. This compounds isolation: employees who use AI without permission fear getting "found out," while those uncertain about policy avoid asking questions.
Mira Greenland, CRO at INTOO, flagged a second-order effect: "When employees outsource their work to AI without oversight or guardrails, deep analysis and collaboration decrease, and imposter syndrome increases." Workers outsource to ChatGPT or Claude to solve problems without immersing themselves in organizational context or validating the output—a shortcut that trades work quality for speed.
The silence is the problem
This isn't a skills gap. Employees adopt aggressively. The friction is institutional: organizations have not drawn clear lines around acceptable use, nor have leaders modeled transparency by admitting their own AI use. The result is shadow adoption with no feedback loop.
Without guardrails, three risks compound: data leakage from unvetted tools, degraded output quality when employees skip analysis, and psychological burden on workers who believe their contributions might not be credited. The last point matters most. Fear of disclosure suppresses both collaboration (the 42% who won't ask for help) and accountability (the 25% who hide their use).
Organizations that silence AI use don't eliminate it. They eliminate visibility into how it's being used and whether it's working.
How to build trust instead of compliance
Greenland outlined three structural moves: establish clear AI guidelines to reduce ambiguity; normalize transparency by having leaders openly document their own AI use; and foster psychological safety so employees feel safe asking questions.
The reframing matters. Don't call it "asking for help"—call it "seeking perspective" or "validating an approach." The highest performers in any domain aren't the ones who never ask questions. They're the ones who ask before small problems harden into bigger ones.
The survey reveals that workers aren't resistant to AI—they're resistant to shame. Close the guidance gap and the adoption you're already seeing accelerates with actual output improvement, not just velocity.