UK senior leaders more likely to use Shadow AI tools

TrustedTech has published research suggesting that UK senior leaders are more likely than junior staff to use unapproved artificial intelligence tools at work. The findings are based on a survey of 2,000 employees in the UK and US.

The study found that 62% of UK senior leaders use so-called Shadow AI tools, compared with 31% of employees below decision-maker level. It also found that 51% of senior leaders are concerned about employees using such tools, despite their own reported behaviour.

Shadow AI refers to unapproved artificial intelligence tools that may process or store company data outside authorised systems. The report says the issue is especially sensitive at senior level because leaders often have broader access to internal systems, customer information and administrative controls.

The research suggests use of unapproved tools is often deliberate rather than accidental. Across UK employees, 22% said they would continue using AI tools even if their organisation banned them and disciplinary action was possible, rising to 28% among senior decision-makers.

The findings point to a gap between formal rules and day-to-day practice as employers try to set limits on generative AI tools. They also suggest some executives see enough benefit in these services to accept the risk of breaching internal policy.

Why staff use them

Among employees who use unapproved AI tools, 24% said limited access to employer-approved products was a factor. Another 21% said unapproved tools were more efficient, while 21% said they did not want their data stored or accessed by their employer.

The survey also points to unease over how AI use is viewed at work. Some 20% of respondents said AI usage could harm career progression, and 19% worried it could raise doubts about their performance or capability.

That concern appears to feed a broader culture of concealment. Some 23% of UK employees said they reduce their own use of AI because they worry about how colleagues or management may perceive them, while 21% said they judge colleagues negatively if they rely heavily on AI to support their workload.

Monitoring is another factor shaping behaviour. The research found that 28% of employees are concerned their employer monitors AI tool usage, adding to distrust around approved systems and potentially pushing staff towards services outside company oversight.

Governance problem

The findings come as businesses try to balance productivity gains from AI tools with concerns over data leakage, compliance and control. Many organisations are introducing approved internal systems alongside policies restricting the use of public tools, but the survey suggests those rules are not being followed consistently.

TrustedTech found that 76% of UK employees acknowledge that using unapproved AI tools poses security or data privacy risks, yet 47% still use them at work. That combination of awareness and continued use suggests risk alone is not enough to change behaviour.

The report also raises questions about leadership example. If senior figures are among the heaviest users of unauthorised tools, companies may find it harder to persuade other employees to follow formal rules or use approved alternatives.

A key finding is the contrast between leaders' concerns about Shadow AI and their own reported use of it.

"Businesses often assume Shadow AI is a bottom-up problem, but our research shows it is being driven from the top down. Senior leaders are not only the biggest users of unapproved AI tools, they are also knowingly bypassing safeguards because the perceived benefits outweigh the risks. When that behaviour is modelled at the top of an organisation, it becomes significantly harder to enforce governance elsewhere in the business. The findings highlight an urgent need for organisations to rethink how they approach AI governance, with a focus on leadership accountability, clearer usage policies, and improved education around secure and responsible AI adoption," said Julian Hamood, founder and chief visionary officer at TrustedTech.

The figures add to a wider debate over whether company AI policies are keeping pace with employee behaviour. While many employers have focused on the possibility of junior staff experimenting with public chatbots and text-generation tools, the survey suggests the challenge may be more entrenched among the people responsible for setting policy.

They also indicate that access remains a practical issue. Where approved tools are limited, slow to roll out or seen as less useful, staff may seek alternatives regardless of policy. For companies, that creates a dual challenge: stopping unauthorised use while making approved options practical enough for daily work.

Overall, the research portrays workplace AI use as shaped by pressure, convenience and mistrust as much as by policy. Nearly half of UK employees surveyed said they still use unapproved AI tools at work.