Annual UK worker spending on AI tools used for jobs is estimated at £1 billion by Deloitte, whose survey also found some employees rated self-funded or potentially unapproved tools above company systems.
Deloitte’s 2026 GenAI Workforce Survey covered 25,000 UK workers, with fieldwork in May and June. In a weighted group of 3,562 people who paid for an AI tool for work or used one that might not be approved, 21% said those tools performed better than company tools. Another 14% in the group said the tools were essential to their jobs but were not funded by their employers. Respondents could select more than one answer.
Deloitte also reported that 17% of GenAI users pay for their own tools. In the 3,562-person group, 16% said the tools could improve their promotion prospects, suggesting some workers see a personal career benefit. The findings do not establish why people used the tools or whether employer limits prompted them to seek alternatives.
In a Nov. 17 opinion column in The National, Alejandro Martin Sposato, an assistant professor in Zayed University’s College of Business in Dubai, recounted an unnamed professional’s complaint that an employer-approved AI platform had an allowance too small for much of the person’s work. Sposato did not identify the worker, employer, platform or limit.
Sposato cited cost, cybersecurity, confidentiality and regulatory obligations as reasons employers may restrict AI access. “A uniform allowance might be straightforward to administer, but it will not suit every role,” he wrote. The account does not establish that usage limits cause workers to use personal or potentially unapproved services.
Deloitte placed employee use of AI outside company channels within the longer history of shadow IT, pointing to earlier workplace adoption of smartphones and cloud software. EconoTimes has previously covered concerns that unapproved AI tools can expose company data.
The UK government’s voluntary AI cybersecurity Code of Practice, published Jan. 31, 2025, advises organizations to encourage employees to report possible AI security risks and to match AI use to the sensitivity of the data involved.



