Australian employees are experiencing a new form of workplace anxiety they could not have envisaged five years ago: AI guilt. They’re grappling with a moral dilemma in which they view AI use as cheating while conceding it enhances the quality of their work.
Almost half the workforce is navigating this conflict, according to the AI Paradox Report, a collection of key insights from almost 9,000 employers and employees in Australia, UK, New Zealand, and Canada.
The cause appears to be rooted in psychology. With technology evolving faster than cultural expectations, deep-seated beliefs around hard work, effort and expertise are being challenged.
Michael Batko has observed the phenomenon in his role as co-founder of Hourglass AI, which tailors AI workflows and creates agents for small and medium businesses. He blames an absence of guardrails.
“The guilt isn’t really about AI. It’s about permission,” he explains. “Employees feel guilty using AI because no one ever gave them explicit permission. People use it in secret and feel dirty about it.”
He believes the solution is to reframe people’s relationship with AI at work, in a process that must begin at the top.
AI Is Redefining Hard Work And That’s Hard To Swallow
42 per cent of people surveyed in the AI Paradox Report say using AI at work feels like cheating. 41 per cent admit to feeling guilty while using it. For 33 per cent, guilt pushes their behaviour underground: they confess they hide their AI use from their employer.
Psychologists and philosophers believe the cause of AI guilt is twofold. Firstly, when change occurs without boundaries, the ensuing uncertainty breeds anxiety. The breakneck pace of AI evolution has left little time for workers to adjust and given leaders limited opportunities to guide the transition. Studies have confirmed uncertainty, or role ambiguity, is a greater psychosocial workplace hazard than overwork or conflicting demands. When AI starts redefining how a role should be performed, that ambiguity creates discomfort.
Secondly, there’s the fact AI has upended the concept of work that people have held for generations. Modern employees have grown up believing valuable work is the result of a struggle. Hustle culture associates success with grinding, and defines the most successful employees as those who put in the longer hours or solve the biggest problems.
When AI produces similar outcomes with a fraction of the effort, that definition of hard work is undermined and people feel their work ethic is in question. “Businesses say, ‘Use AI’ but the day-to-day signal employees actually read is, ‘Don’t get caught looking lazy,’” Batko says. The concern is warranted: an Atlassian study found people considered a hypothetical peer who disclosed AI use 10 times lazier than one who didn’t.
Employees are also conflicted by the concept of ownership. People who have built a career on reputation and expertise can find it hard to take credit for work that is not entirely theirs. There is also ‘anticipatory guilt,’ where people fear being judged as inauthentic or less capable if they disclose their AI use. This concern is valid, with studies confirming people are less likely to trust AI-generated information than human-led communication.
Despite underlying discomfort, employees are not turning their backs on the technology. AI Paradox report data shows 59 per cent of users believe AI is helping them build valuable skills while almost three-quarters agree it has directly improved the quality of their work and boosted their daily productivity.
Businesses Lack AI Oversight And Instruction
Uncontrolled AI use is more common in small and medium businesses than in larger organisations, possibly due to the fact the latter tend to have more resources and clearer guidelines.
The issue goes beyond secrecy. 44 per cent of employees log in to personal AI accounts in their workplace. This opens the door to unauthorised use of proprietary data and exposes small and medium businesses to risk.
Businesses without AI governance policies are also more likely to lack structured training. Half of workers surveyed in the AI Paradox report say their employer provides little or no AI education, despite business owners saying they favour AI skills in new hires. To fill the void, more than half of employees are training themselves via social media platforms, including TikTok, YouTube, Reddit, and Discord. Many don’t mind: 41 per cent view AI upskilling as their own responsibility, not their employers’, particularly Gen Z workers.
It’s not only employees learning how to use AI. Many employers are also trying to get their heads around the new technology. Batko recommends they ignore social media influencers and look to real-world examples.
“The best teachers aren’t the loud voices online,” he says. “They’re the founders and operators actually using it every single day. Look to the bookkeeper down the road who quietly automated their month-end.”
But no matter how proficient staff become in AI use, he reminds leaders of the value of human oversight, especially when it comes to payroll and accounting. “Let AI do the first 80 per cent: the drafting, sorting, summarising. Keep human judgment on the 20 per cent that involves someone’s pay or a compliance deadline,” he recommends.
Leaders Have The Power To Quash AI Guilt
Eradicating AI guilt requires leaders to remove the uncertainty that enables it, says Batko. The fix lies in explicit permissions, and clear guidelines and governance policies, with business owners leading by example. Employees should be left in no doubt as to when AI should and should not be used.
“If leadership wants AI used, they have to say so out loud and use it themselves first,” he insists. “The exec team has to use it out loud, in front of everyone, and say, ‘This is how we work now.’ The guilt evaporates the moment leadership goes first.”
Batko recommends not only bringing AI usage into the open but celebrating it. He suggests running regular “What did AI do for you?” sessions, where people swap examples of efficiencies and time savings the technology has enabled.
“Spend five minutes a week sharing one thing AI did for you,” he says. “It turns a guilty secret into a team habit.”
It can also promote growth. When employees are using AI in secret or in silos, they are less likely to share efficiencies or improved workflows with others and can waste time by independently trying to solve the same problems. There is also the potential for stifled innovation when people are reluctant to share ideas brainstormed via AI.
Leaders Must Tailor AI Adoption To Their Businesses
Batko has practical advice for small business owners keen to close the permission gap and formalise AI use within their organisations. While he advocates for roadmaps, he says it’s too easy for leaders to get caught up obsessing over strategies and plans that are never implemented.
The solution, he says, is to roll up their sleeves. “Start with one painful hour,” he says. “Don’t try to transform the business. Just find the single most annoying recurring task this week and point AI at just that. Momentum beats a master plan.”
He also believes in targeted adoption, based on need rather than software. “Don’t buy the shiny chatbot. Fix the workflow,” he says. This means integrating where it will be useful, not where it looks impressive. “Most small businesses don’t need a flashy AI front-end. They need the boring, repetitive grind in the back office automated and taken off their team’s plate.”
He adds that one-size-fits-all approaches to AI adoption are unlikely to set up businesses for growth. “Everyone has access to the same AI tools. Your edge is your own messy data: the processes and knowledge trapped in your team’s heads.” He describes this as a small business’s moat. “For an accounting or payroll firm, the gold is your actual client docs, checklists and how-we-do-it-here knowledge. That’s what makes AI useful instead of generic.”
This doesn’t mean generic tools can’t be optimised. Batko says without training, staff are unlikely to get more than a surface-level benefit from large language models like Claude, Gemini, and ChatGPT.
“Stop treating AI like Google,” he says. “It’s not a search engine. The magic isn’t in the question, it’s in the context. Give it your real templates, your tone, your process, and it stops being generic.” A better option would be to regard AI assistants as promising interns.
“Treat it like a brilliant new hire: give it background, show it how you work and delegate properly.”
While a reduction in guilt is harder to quantify, Batko says business owners can track time savings to see whether AI integration is paying off. Indeed, 65 per cent of employers surveyed in the AI Paradox report say the technology is accelerating their business and allowing them to work faster. Nearly three-quarters of employees have increased productivity and 87 per cent of expert AI users are diverting time saved to higher-impact tasks.
“The only AI metric that matters for a small business is hours back in your team’s week,” he maintains. “If you can’t feel it, AI is a toy.”
