Nearly three in four UK employees are quietly deploying shadow AI tools in the workplace, a new Microsoft-commissioned survey reveals. The study, conducted via Censuswide across over 2,000 workers in various sectors, found that 71 % admitted to using AI systems not officially approved by their employers.
The survey highlights that more than 40 % of workers adopt these unapproved AI systems because they already use them in their personal lives, making integration into workflow seamless. Common applications include responding to emails and internal communications, drafting reports and presentations, and—less commonly—managing finance-related tasks.
Productivity Gains vs. Rising Risks
Users of shadow AI report saving nearly eight hours per week on average, mostly by automating administrative tasks. Microsoft extrapolates this to an estimated 12.1 billion hours saved annually across the UK economy, equivalent to around £208 billion in workers’ time.
However, these gains come with serious downsides. Microsoft warns that unapproved AI use exposes companies to data leaks, regulatory non-compliance, and increased risk of cyberattacks. Only about 30 % of respondents expressed concern over privacy or company data exposure when using these tools. Darren Hardman, CEO of Microsoft UK & Ireland, emphasised that “AI tools should be built for the workplace, not just the living room,” noting that only enterprise-grade AI can offer the security and privacy protections organisations require.
Recent Trends and The Bigger Picture
The rise in shadow AI comes amid increasing urgency for businesses to adopt AI responsibly. Earlier reports show that workplace AI usage in the UK surged by 66 % since late 2023, especially in non-tech roles, where AI skills are now pervasive.
Still, many organisations lag in providing safe, approved tools: according to a YouGov study, just 31 % of UK SMEs currently leverage AI solutions, citing concerns over data privacy and ethics as primary barriers.
Regulators are also catching up. The UK government launched the AI Safety Institute following the 2023 AI Safety Summit, and continues to balance innovation incentives with governance frameworks. The use of shadow AI puts pressure on firms to accelerate adoption of controlled, compliant, and secure AI systems—or risk being blindsided.
Moving Forward: How Organisations Can Respond
To mitigate shadow AI risks, companies should:
•Provide sanctioned, enterprise-grade AI tools to employees
•Deliver training and clear guidelines on responsible use
•Monitor AI usage and establish governance protocols
•Integrate security and data protection features into AI deployments
By proactively managing AI adoption, organisations can harness productivity benefits while safeguarding their data and compliance posture.
