40% of Aussie Workers Feel Guilty Using AI—What It Means for Tool Design

AI developers face a fascinating user behaviour puzzle: why do two in five workers feel guilty about using a tool that demonstrably improves their output? The new ‘AI Paradox at Work’ data suggests that current chatbot experiences lack the social and ethical cues that make augmentation feel legitimate. This is a design imperative—future AI assistants must embed features that normalise augmentation, offer clear disclosure mechanics, and integrate with workplace identity to remove the ‘shadow user’ stigma.

A new study has revealed a striking paradox in Australia’s workforce: while three-quarters of workers say artificial intelligence tools significantly boost their productivity and work quality, one in three are using them in secret—without telling their employer. The findings, released by Employment Hero in its ‘AI Paradox at Work’ report on 7 July 2026, are based on a survey of over 1,000 business leaders and 1,600 employees, and paint a picture of an underground AI revolution fuelled by guilt, self-training, and a lack of organisational support.

“

Yet, 33% admit to using tools like ChatGPT or other chatbots covertly, with 44% of employers reporting that staff are accessing personal AI accounts at work.

The data shows that 75% of Australian workers believe AI has improved their productivity, and 74% say it has enhanced the quality of their output. Yet, 33% admit to using tools like ChatGPT or other chatbots covertly, with 44% of employers reporting that staff are accessing personal AI accounts at work. The secretive usage is compounded by a pervasive sense of cheating: two in five employees said using AI for work feels like cheating, a sentiment that Employment Hero’s Asia Pacific managing director James Keene says often leads workers to hide their activities.

This phenomenon unfolds against a broader backdrop of Australian AI adoption. The National AI Centre recently reported that almost half of Australian organisations are now adopting AI technology, but 19% remain unsure how to use it effectively. The gap between individual worker experimentation and formal enterprise strategy is widening. Workers are overwhelmingly self-educating—over half are turning to TikTok, YouTube, and online blogs rather than company-provided training. Keene described it as a “pretty interesting place” where employees are forced to seek external resources just to be more productive for the businesses they serve, a scenario unimaginable five years ago.