By Nicky Marks, Managing Director
In the less than three years since ChatGPT launched, artificial intelligence has achieved something remarkable: near-universal acceptance among business leaders. Our latest research reveals that 93% of business leaders believe AI should be integrated into their teams’ work — one of the fastest technology adoptions in modern workplace history.
But beneath this enthusiasm lies a contradiction that could undermine the very productivity gains leaders are chasing.
The self-teach paradox
While business leaders are championing AI adoption at unprecedented speed, they’re often failing to provide their teams with the tools to succeed. Our survey of over 1,000 business decision-makers uncovered a stark reality: only half (49.7%) provide structured training to their employees, despite nearly universal support for AI integration.
Even more concerning, 64% of leaders expect their team members to train themselves on AI tools. It’s a “figure it out yourself” approach to what may be the most transformative workplace technology of our generation.
This creates what we’re terming the “self-teach paradox” — leaders are moving five times faster on AI adoption than on training implementation. The gap between expectation and support represents a real vulnerability for organisations racing to stay competitive.
A generational earthquake
Perhaps the most dramatic finding in our research is the wide generational divide reshaping workplace dynamics.
Among leaders aged 18-24, an overwhelming 91.7% support extensive or significant AI use. That number plummets to just 23.3% among leaders over 65 — a 68-point difference. Younger managers are four times more likely to champion extensive AI integration than their older counterparts.
As younger, AI-native leaders rise through organisations, it’s likely that they’ll increasingly clash with senior leadership who view the technology with skepticism. The question isn’t whether this generational divide will cause friction, but how organisations will manage it.
Leaders practice what they preach (sort of)
To their credit, business leaders are using AI tools themselves. Nearly one in three (29.7%) use AI daily or more frequently, with 67.4% using AI at least weekly. This represents faster personal adoption than we saw with previous workplace technologies like smartphones or cloud computing.
But things get complicated when leaders say they want employees to feel confident using AI (72%), yet 52% admit it bothers them when they notice team members actually using it. They believe AI has an important function in workplace communication (74%), yet nearly 70% claim they can detect when AI has been used in their team’s work.
We’re calling this the “AI transparency paradox” — a contradiction that risks creating a toxic “don’t ask, don’t tell” culture around AI usage. Employees receive mixed signals: use AI to be more productive, but don’t let me notice you’re using it.
What keeps leaders up at night?
Data security tops the list of concerns at 40%, followed by over-reliance on technology (34.9%) and reduced human connection (32%). Interestingly, job displacement ranked lowest at just 21.2%, suggesting leaders either don’t believe AI will displace workers or view this as an acceptable tradeoff.
Despite these concerns, optimism prevails. Looking ahead five years, 61% of leaders believe AI will improve work quality or empower employees to do more. Plans for expansion are aggressive: 69% intend to increase their own AI use in the next year, while 66.5% plan to boost AI adoption across all employees.
The path forward
This research reveals a pivotal moment in workplace evolution. Organisations that bridge the training gap and establish clear, supportive AI policies will gain significant competitive advantages. Those that maintain the current paradox—demanding AI adoption while providing inadequate support and sending mixed signals about usage—risk creating anxiety, inefficiency, and a culture of dishonesty.
The race to integrate AI is already won. The real competition now is about who can do it most effectively, most transparently, and most humanely. The data suggests we have work to do.
This research was conducted by Censuswide among 1,002 business decision-makers between October 21-22, 2025.
