Eight in 10 UK workers are now using artificial intelligence to support their day to day tasks, but a majority are choosing not to tell managers or colleagues, according to new research.
It suggests a pattern of quiet adoption across workplaces where employees rely on AI tools for speed and accuracy but feel uncertain about whether those tools are officially supported.
The study, carried out by learning technology firm Cornerstone among 2,000 UK employees, shows 63 percent use AI weekly or more. Of those who do, 81 percent said they did not always disclose their use. Many said they felt embarrassed or unsure about using the technology openly, and some pointed to fears that others might see AI as a way of cutting corners. The data indicate that employees are adopting AI on their own initiative while waiting for clearer direction from employers.
The secrecy around AI sits alongside a widening training gap. Half of employees said they had never or rarely received training or tools to support AI use, and just eight percent said they always received formal support.
 |
Get our essential daily HR news and updates. |
Hidden use drives quiet experimentation
The research found that 62 percent of AI users felt embarrassed or ashamed about using the technology at work. Some worried about being judged. Others said colleagues might assume they were not capable of completing tasks without assistance. At the same time, a significant group did not share these concerns. Thirty eight percent said they never felt embarrassed and noted that AI saved time, improved accuracy and helped with creativity.
The mix of confidence and uncertainty suggests a workforce moving at different speeds. Employees who feel comfortable with the technology are using it to streamline workloads.
Those who feel uneasy are continuing to use AI quietly, often without guidance. Without clear policies, experts say it may lead to inconsistent practices across teams and a lack of shared understanding about how AI should be used.
Inequality in access to training
The data show notable differences between salary groups. Twenty four percent of workers earning under £25,000 said they had never received AI training. Among those earning £55,001 or more, the figure fell to 13 percent. This indicates that those in junior or lower paid roles may be missing out on development opportunities that could help them navigate new workplace tools.
There is also strong interest in learning. In the lowest pay band, 46 percent said they were positive about receiving AI training. In the highest pay band, the figure rose to eighty six percent. This suggests employees want to build confidence and use AI responsibly if given proper support.
The findings also highlight the risk that frontline or so called deskless workers, who do not have regular access to devices or training opportunities, could fall behind, and that employers that do not address this gap may face growing inequalities within teams.
Managerial silence reinforces uncertainty
Only one third of employees said they were often or always encouraged to use AI at work. The absence of clear direction may lead staff to keep their use private. Some may use AI at home and then bring results into work without explaining how they reached those outcomes. This can affect quality checks, collaboration and trust between colleagues.
Workplaces that have not updated policies or offered training may be contributing to the secrecy. Employees who are unsure about expectations are more likely to experiment quietly rather than ask for guidance, potentially leaving organisations unaware of how AI is being used day to day.
Dominic Holmes of Cornerstone said lower income and deskless workers were often left behind without structured training. “It’s important that lower income workers, who are typically more skill based, receive the training they deserve,” he said. “The ‘deskless’ workforce should also have access to the same AI training as office based employees. Without this formal support, lower income workers are often left behind, relying on AI informally and without organisational guidance creating a clear AI skills gap.”
He added that “[b]usinesses that invest in structured training and open dialogue around AI for all salaries, will not only bridge the skills gap but also build a more confident, transparent, and future ready workforce”.
Holmes’ remarks reflect the broader concern that rapid AI adoption may widen existing workplace divides if employers do not intervene. When staff receive uneven levels of support, gaps in confidence and capability grow over time.
The study, experts say, indicates that employees want clarity on when and how they should use AI. Many said the technology helps them complete tasks more efficiently and improve the quality of their work. But without training, they may rely on trial and error or copy practices from colleagues who are also unsure.
