63% of UK Workers Now Use AI at Work, 10% Violate Company Policy

Unauthorized AI use spreads across UK workplaces

Nearly one in ten UK workers have used an AI tool their employer has either banned or would disapprove of, according to a new survey from Deloitte that underscores mounting friction between employee adoption and organizational policy.

The survey of 25,000 respondents found that 63 percent of workers have used AI tools at work, suggesting strong appetite for technology that can automate routine tasks or enhance productivity. Yet the same proportion expressed concern that their manager might conclude AI could replace them entirely.

Paul Lee, who led the survey, said workers are simultaneously thinking "this could increase my profitability or this could diminish my worth," creating what he described as a "stigma" around workplace AI adoption.

Security risks of shadow AI

The phenomenon of employees using unsanctioned AI tools—often called "shadow AI"—carries significant security implications. Data security firm Cyberhaven, which monitors data flows across hundreds of companies, found that one-third of employees access AI through personal accounts, with nearly 40 percent of AI interactions involving sensitive information. A separate Microsoft report in October placed unauthorized AI usage at 71 percent.

Last week the UK National Cyber Security Centre issued a warning about shadow AI risks, including exposure of confidential data and loss of decision-making transparency. David Chismon of the NCSC said the practice is "widespread" and urged IT security teams not to assume they have full visibility.

"Organizations can't hope to block connections to all possible AI tools, so they need to develop a positive cyber security culture with open dialogue about the tools staff might wish to use and to set clear guardrails around what secure use of AI looks like," Chismon said.

James Longster, a technology partner at law firm Travers Smith, compared the trend to earlier unauthorized IT practices involving USB drives and personal cloud storage. He warned that consumer AI versions typically ingest data for training purposes: "An employee might be uploading the crown jewels into a commercially available AI."

Enforcement remains challenging

Gloucestershire Police has dismissed a staff member for using AI against force rules and circumventing blocked sites through personal accounts. Separately, the Independent Office for Police Conduct is investigating an officer for alleged gross misconduct related to AI use.

Ailie Murray, an employment partner at Travers Smith, said businesses are increasingly concerned but enforcement is difficult. "Employers don't necessarily have the evidence if employees are using, for example, their own personal ChatGPT on their mobiles," she noted, adding uncertainty about whether personal AI use constitutes general knowledge building or actual work completion.

Why it matters

The gap between employee AI adoption and organizational readiness creates both security vulnerabilities and missed productivity opportunities. Most Deloitte survey respondents said company leaders have poor or nonexistent understanding of AI, while half of AI users received no formal guidance or training. As the UK Office for National Statistics cites AI as a major driver behind the country's 2.4 percent annualized growth rate, organizations that fail to establish clear AI policies and training risk both data exposure and competitive disadvantage.

The findings were first reported by Claims Journal, citing Bloomberg coverage.