Rampant ‘shadow AI’ use in UK workplaces raises concerns: Microsoft
WASHINGTON, UNITED STATES — Seventy-one percent of United Kingdom employees are routinely using unapproved consumer AI tools for work tasks, creating significant security risks despite generating substantial productivity gains, according to Microsoft research.
The study, commissioned by Microsoft and conducted by Censuswide in October 2025, found that the majority of workers have used this “shadow AI,” with 51% continuing to do so weekly, saving an estimated 12 billion hours annually but often without regard for data privacy.
This research is based on a nationally representative survey of 2,003 UK employees, which included minimum quotas of 100 respondents each from the Financial Services, Retail, Consumer Goods, Education, and Health & Social Care sectors, as well as 500 each from large businesses and public sector organizations.
“UK workers are embracing AI like never before, unlocking new levels of productivity and creativity. But enthusiasm alone isn’t enough,” said Darren Hardman, Chief Executive Officer (CEO) of Microsoft UK and Ireland.
Pervasiveness and peril of shadow AI
They are mostly employed in core activities: 49% on drafting and responding to communications, 40% on developing reports and presentations, and 22% on finance-related activities, indicating that workers heavily depend on external platforms to conduct sensitive business activities.
This reliance creates substantial vulnerability, as employee awareness of the risks is alarmingly low. Only 32% of respondents expressed concern about the privacy of company or customer data input into these tools, and a mere 29% were worried about the security of their organization’s IT systems.
The motivations of this behavior include comfort and familiarity, where 41% say they use the tools that they are used to with personal lives, and 28% say that no option is provided by the organization, which exposes organizations to possible data breaches and cyberattacks despite the productivity gains.
“The message is clear: only enterprise-grade AI delivers the functionality that employees want, wrapped in the privacy and security every organization demands,” Hardman notes.
Productivity gains and a cultural shift
The undeniable power of AI tools lies in the compelling efficiency they deliver, which is why they are rapidly adopted without authorization.
Users of generative AI at their workplaces say that they save 7.75 hours a week on administrative work, which would save 12.1 billion hours annually across the UK economy, worth around 208 billion.
Employees are using this saved time to enhance their work-life balance (37%), acquire new skills (31%), and do work that matters to them (28%), which is showing a tangible, positive influence on their work experiences.
This productivity increase is driving a broader cultural shift towards the acceptance and establishment of AI in the UK labor market. This has greatly improved employee mood, as half of them have become optimistic, excited, or confident about AI —a vast change from January of this year, when the figure stood at 34%.
Besides, perceptions of AI’s strategic value have more than doubled, with 39% of workers now considering it a critical component of organizational success, up from 18% at the beginning of the year, indicating a shift towards the mainstream legitimization of AI in the business domain.
This widespread, unapproved adoption of AI forges a dangerous reality for business, where the urgent need to harness profound productivity gains is now inextricably tied to the critical task of mitigating security vulnerabilities.
