Workers are lying less about AI—but they’ve developed a worse habit

A recent survey by WalkMe reveals that despite worker confidence in AI usage, many are overestimating its effectiveness. The survey found that the number of employees admitting to using AI meaningfully in the workplace has dropped significantly, from 8% to just 2%. While fewer people are lying about their AI abilities, the survey suggests that workers are becoming more comfortable with the technology, rather than necessarily becoming more skilled.

According to WalkMe cofounder and CEO Dan Adika, the confidence gap between employee perceptions and actual results is troubling. Adika argues that the mathematical reality does not support the high confidence levels reported in the survey. He believes that companies are not seeing the anticipated savings in hours and dollars from AI implementation, despite workers feeling confident in using it.

The survey also highlights a significant disconnect between employees' perceptions of AI and the reality of its implementation. Employees feel that AI can save time, but they are struggling to see tangible business results. Additionally, the survey reveals that employees feel a senior leader may not fully understand the AI strategy being championed, as 53.6% of respondents expressed this concern.

Adika warns that this overconfidence in AI can be dangerous, as employees may be asked to pour their expertise into AI systems while ultimately becoming less essential. He likens the situation to creating a "company brain" that puts employees at risk of being replaced. The survey suggests that the issue lies more in the understanding and acceptance of AI's limitations than in deliberate corporate schemes.

Adika advocates for more targeted training and better integration of AI tools with existing systems, rather than simply providing more training to employees. He also highlights the importance of addressing permission structures and other technical challenges that hinder AI implementation in real-world scenarios.