A major study into AI use in the workplace found that most workers surveyed aren’t completely honest with their bosses and colleagues about how they use it.
KPMG’s Trust in AI report, which it carried out with the University of Melbourne, surveyed 48,340 people across 47 countries between November 2024 and January 2025.
It found that more than half of employees asked, 57%, admitted to hiding their use of AI at work and present AI-generated content as their own.
KPMG says the report, titled “Trust, attitudes, and use of artificial intelligence: a Global Study 2025,” is one of the most comprehensive looks at attitudes towards AI in the workplace to date.
It shows just how embedded AI has become in the modern working world — 58% of people now intentionally use AI for work, and about a third do so at least once a week, the study found.
Nicole Gillespie, one of the study’s authors and a professor of management and chair of trust at the University of Melbourne’s business school, said there is a surprising level of “inappropriate, complex, and non-transparent” activity in how employees use AI.
Covert use of AI at work is happening because there’s pressure to use the tools and keep up with others’ progress, Gillespie said. Employees feel that unless they start using these new tools, they’ll be left behind, putting their jobs at risk, she said.
If employers ban the use of Gen AI, employees have to hide what they’re doing. But Gillespie added that there’s also “a seductive element” to why they’re not owning up to using AI tools.
She said that once people start seeing the benefits, it’s tempting to keep using AI, even if they know they’re breaking company policies by doing so.
Hiding how you use AI creates ‘significant risk’
Only 47% of employees globally say they have received AI training, the study found, which means many are just taking it into their own hands and using the technology without guidance.
The data wasn’t positive about their capabilities: 66% of employees said they use AI tools without evaluating the accuracy of responses, 48% said they have uploaded company information into public AI tools, and 56% reported making mistakes in their work due to AI.
“That is really concerning because that’s where the organization is exposed to significant risk,” said Sam Gloede, the global trusted AI transformation leader at KPMG International, who worked with Gillespie on the study.
Gloede said that in addition to errors, data breaches, and compliance risks, a big concern is that the lack of transparency around AI use in the workplace will reinforce distrust in AI systems.
Trust is a strategic asset for organizations, allowing them to innovate and grow, said Gloede. When it comes to AI, users need to trust that the technical capabilities are there, that it’s fit for purpose, and it’s reliable, she said.
“So much of what we do both in organizations, but more broadly in society is underpinned by trust,” added Gillespie, who is a world-renowned expert in organizational trust.
“In the last five to 10 years, it’s become a centerpiece of pretty much every board. They recognize the importance of not only building but also preserving stakeholder trust in the organization,” said Gillespie.
Better AI literacy leads to better outcomes
To build trust around AI and get accurate and effective results, organizations have to provide better AI training and governance, the two experts said.
The study found that literacy and governance are lagging in workplaces, with half of those surveyed saying they don’t understand AI and how it’s used. Only 2 in 5 have reported any AI-related training or education.
Gillespie said employees need both foundational training on what AI is and its ethical and responsible use, and role-based training so that everyone understands how to use the technology to increase productivity and avoid errors.
“If we really want that deep education, it has to go beyond just the training. It’s also about helping people to be using AI in the workplace transparently, to be using it openly, to be sharing their learning, to have communities of practice, to have structures set up where they can experiment and learn and make mistakes without it creating risks,” she said.
In emerging economies like India, Nigeria, and Saudi Arabia, trust levels in AI were much higher than in advanced economies — 82% compared to 65%.
Gloede highlighted that this also correlated with higher literacy and training levels. “It really calls out that the literacy and the training is a vital ingredient to improve that adoption and success,” she said.
Have a tip? Contact this reporter via email at pthompson@businessinsider.com or Signal at Polly_Thompson.89. Use a personal email address and a nonwork device; here’s our guide to sharing information securely.
Read the full article here
