How to Make Your Workplace Culture Great with AI

By Dr Anna Kiaos

Most organisations are asking how to adopt AI without damaging their culture. That is the wrong question. The better one: how do you use the arrival of AI to enhance your workplace culture(s) more than before?

A third of Australian workers are using AI at work without telling their employer. More than two in five say that using it feels like cheating. The same workers report that AI is making them more productive and lifting the quality of what they produce. Those findings come from Employment Hero's AI Paradox at Work study, released this week — a survey of more than 5,400 workers and 3,200 employers across Australia, Canada, New Zealand and the United Kingdom — which I was engaged to help interpret as an independent expert. The national coverage that followed, fixed on the secrecy. Understandably: "workers hiding AI" is an arresting headline.

In my view, secrecy is the symptom, not the condition. The condition is cultural. Read the way an organisational ethnographer reads it, the data carries something hopeful: the raw material for enhances workplace culture(s) is already inside your organisation. It is simply sitting backstage.

Shadow AI is backstage behaviour

Erving Goffman taught us that organisational life runs on two stages. The frontstage is where employees perform for an audience — leaders, systems, formal expectations. The backstage is where they behave as they actually are: candid, improvisational, unguarded.

Hidden AI use is backstage behaviour in its purest form. On the frontstage, employees comply with what they believe the organisation expects — which, absent any clear guidance, they read as "don't use it." Backstage, they experiment, trade prompts with trusted colleagues, and quietly redesign their own jobs. The study suggests most workers using AI taught themselves — through TikTok, YouTube and blogs, not through anything their employer gave them.

Workers aren't hiding AI because they're doing something wrong. They're hiding it because nobody has told them what right looks like.

In my earlier writing on workplace microcultures, I described how groups form beneath the official culture as a protective response when employees feel the formal culture no longer reflects their reality. Shadow AI use runs the same way. When an organisation stays silent, employees don't stop using AI — they stop talking about it. Silence doesn't remove the behaviour; it drives it underground, where you can neither guide it, learn from it, nor manage its risks.

The guilt is the signal

The most revealing number in the Australian data isn't the secrecy. It's the shame. When two in five workers say that using a sanctioned productivity tool feels like cheating, they are telling you how effort and worth are wired into your culture.

Guilt of this kind is an identity signal. It says: I no longer know whether my work counts as mine. For professionals whose sense of self is built on craft and competence, AI reaches past workflow and into the story they tell themselves about their own value. Organisations that treat this as a compliance problem — "disclose your AI use" — will miss it completely. It is a psychosocial problem, and it answers to what psychosocial problems always answer to: clarity, safety and leadership.

Five practices that turn AI into a culture-builder

The organisations getting this right aren't the ones with the most sophisticated tools. They're the ones doing the cultural work and creating deliberate open, transparent dialogue with staff. Five practices matter most.

1. Replace silence with explicit permission. A policy is the floor, not the ceiling. People need to hear a leader say, in plain words, "we want you using these tools — here's where, here's where not, and here's who to ask." Until permission is spoken, silence reads as prohibition, and the behaviour stays backstage.

   
   

2. Bring AI use frontstage. Make sharing ordinary: prompt libraries, show-and-tell sessions, "how I automated this" in the team meeting. The moment your best AI users are recognised rather than suspected, secrecy loses its point.

   
   

3. Treat guilt as data, not deviance. Ask your people how AI use feels, not only whether they've declared it. Wherever "it feels like cheating" surfaces, you've found an effort-to-worth story that needs rewriting — and it's the leader's job to redefine valued work once the typing is no longer the point.

   
   

4. Curate the learning already happening. Your workforce is training itself on AI right now. That's initiative worth honouring and worth guiding — vetted pathways, peer mentors, protected time to practise, so capability builds.

   
   

5. Measure beneath the engagement survey. Stable engagement scores can sit on top of a fracturing culture: compliance on the frontstage, workarounds behind it. That gap is precisely what the Culture Pressure Map™ was built to see — where pressure is accumulating, where microcultures are forming, and where the official culture and the lived culture have quietly parted company.

What "great" actually looks like

A great AI culture isn't only one where everyone uses AI. It's one where nobody has to hide anything. Where a junior can say "I used AI for the first draft" in the same breath as "and here's my thinking on it," and both raise their standing. Where leaders model their own use, misfires included. Where feelings of guilt or anxiety — which are real, and rational — are spoken about openly instead of managed silently by each person at their own desk.

The productivity gains are already here. The skills are already growing. The only thing still backstage is the conversation. The organisations with the nerve to bring it forward will find that AI, handled as a cultural transition rather than a technology rollout, hands leaders something unexpected: the most honest look at their culture they've had in years, and a rare reason to make it better.

References and further reading

Employment Hero, The Australian AI Paradox: How Workers Really Feel (2026) — https://employmenthero.com/blog/australian-ai-paradox/

AI on the sly: more Aussies are using chatbots for work, AAP, syndicated via Australian Community Media (2026) — https://www.newcastleherald.com.au/story/9305833/ai-on-the-sly-more-aussies-are-using-chatbots-for-work/

Guilty secret? One in three Aussies using AI chatbots at work, SmartCompany (2026) — https://www.smartcompany.com.au/artificial-intelligence/guilty-secret-one-in-three-aussies-using-ai-chatbots-at-work/

Kiaos, A., When AI Meets the Backstage: How Artificial Intelligence Is Accelerating the Formation of Workplace Microcultures, Mind Culture Life — https://mindculturelife.com.au