23% of Bosses May Overstate Their Own AI Skills, Survey Finds

What happened

A new workplace survey reported by TechRadar finds that 23% of bosses may be overstating their own artificial intelligence skills, even as these same leaders increasingly prize AI literacy in their teams. Almost half of managers surveyed said they now value demonstrable AI skills more highly than formal qualifications when assessing employees.

The findings point to a widening gap between the AI competence organisations say they want to build and the AI competence leaders actually possess, with a meaningful share of managers appearing to exaggerate their own understanding of the technology while setting the bar for others.

Why it matters

This is fundamentally a story about how organisations are managing the AI skills transition, not just a workplace-trust anecdote. As companies lean on AI adoption to reshape roles, hiring criteria and performance expectations, credibility at the top becomes a structural issue: if leaders cannot accurately model the competence they demand, training investments, hiring decisions and AI governance built on that leadership signal are all weaker than they appear.

For transformation leaders, the finding is a useful early-warning signal. Skills gaps are usually framed as a frontline or mid-management problem to be solved with training budgets. This data suggests the gap may start higher up the chain — and that AI maturity programmes which assume competent sponsorship at leadership level could be building on shakier foundations than planned.

By the numbers

• 23% of bosses surveyed may be misrepresenting the extent of their own AI knowledge or skills.
• Almost half of managers say they now rate AI skills above formal qualifications when evaluating talent.

The Renascence take

The headline risk here isn't dishonesty for its own sake — it's what happens downstream when unverified confidence becomes the basis for hiring criteria, training priorities and AI rollout decisions. Behavioural economics has a name for this: competence signalling under social pressure, where admitting "I don't fully understand this" feels costlier than quietly overstating fluency, especially from a position of seniority.

Most organisations treat AI upskilling as a bottom-up problem — train the workforce, measure adoption, celebrate the dashboard. This data suggests the more urgent diagnostic question is upward, not downward: do the people setting AI expectations actually understand what they're asking for? A customer-obsessed, evidence-led operator should treat leadership AI literacy as a measurable capability, not an assumed one — with the same rigour, testing and psychological safety to say "I don't know yet" that they'd expect from any frontline hire learning a new system. Skip that step, and the AI strategy is being steered by confidence rather than competence.