Survey shows workers admit less AI bluffing but grow overconfident

WalkMe’s latest AI at Work Pulse Survey, conducted by Propeller Insights among 2,037 U.S. workers, reveals a decline in admitted AI bluffing—pretending to know AI in meetings fell from 45.2% to 28.3%, and claiming AI-generated work as one’s own dropped from 48.7% to 32.5%. Yet a large share of employees still report feeling confident with AI, while only 24.6% say it succeeds on the first try, and many spend more time troubleshooting than the original task would require.

Dan Adika, WalkMe’s co-founder and CEO, says this confidence often translates into overconfidence, with little evidence of measurable cost savings or clear business impact. He also highlights a double standard where AI errors are scrutinized more heavily than human mistakes, and notes that senior leaders are perceived as lacking full understanding of the AI strategies they promote. The survey points to generational gaps, with Gen Z showing the highest confidence but also the highest rate of overstating skills, leading to workplace problems.