Two thirds of UK workers are now using artificial intelligence tools as a standard part of their working day, yet the headline adoption figure conceals a fault line that matters far more to HR teams. Research commissioned by Adobe Acrobat and surveying 2,000 UK adults between 18 and 22 April found that 68% of workers use AI at work, with nearly two-thirds of those users reporting time savings of up to eight hours per week. Beneath that figure sits a 54-point gap between the C-suite and frontline staff on whether they feel safe sharing work documents with AI at all. That gap is where the real challenge for UK employers begins.

What the Adobe survey measured

Adobe Acrobat surveyed 2,000 UK adults between 18 and 22 April, and the findings were reported by HR Review in late October 2026. The survey concentrated on practical usage: which tasks workers delegate to AI, and what obstacles get in the way.

Document creation and editing leads adoption at 27% of respondents, covering drafting emails, proofreading, and summarising reports. Data analysis and technical tasks such as coding and spreadsheets follow at 23%, with note-taking, brainstorming, and scheduling at 22%. Use of AI for HR and operations stands at 14% overall, rising to 24% among property and real estate workers, where recruitment and onboarding applications are already embedding AI tools.

One methodological note worth keeping in mind: the survey is vendor-commissioned and relies on self-reported time savings. The Office for National Statistics (ONS), which tracks labour productivity through the Annual Survey of Hours and Earnings (ASHE), has not yet published a matched dataset on AI-attributable time savings. The eight-hours figure tells you AI is creating meaningful capacity for many workers. It does not specify, on its own, how those hours are then redistributed, or whether they translate into reduced pressure or higher output targets.

The trust gap: seniority, age, and sector

74% of C-suite leaders say they would be comfortable sharing work-related documents with AI tools, compared with just 20% of employees without management responsibilities. That is a 54-point gap, and it matters because leadership comfort with AI tends to set the speed and ambition of rollout plans. When those plans assume a staff population with comparable confidence, they are built on faulty ground.

Age produces a similar gradient. 52% of workers aged 25 to 34 are comfortable sharing work documents with AI systems, against only 15% of those aged 55 and over. In IT, technology, and telecoms, comfort levels reach 73%, broadly in line with C-suite responses, suggesting sector experience with digital tooling is a stronger predictor of AI confidence than seniority alone.

The gender dimension carries its own implications. Women are more likely than men to report discomfort sharing sensitive documents with AI: 45% compared with 33%. Under the Equality Act 2010, employers have a broad duty to consider whether workplace policies and practices risk producing disparate outcomes across protected characteristics. Where AI adoption rates diverge significantly along gender lines, the risk is not just productivity loss. Over time, uneven adoption can create uneven recognition for output gains, which feeds through into pay and promotion decisions that the Equality Act 2010 is designed to scrutinise.

What is holding workers back

The barriers the survey identified cluster around governance and guidance rather than technical ability or resistance to change:

• Data security and privacy: the top obstacle, cited by 22% of respondents.
• Accuracy of AI output: a concern for 20%.
• Inadequate training or guidance: named by 19%.
• Restrictive company policies on AI use: 14%.
• Not knowing which AI tools to use: 12%.

The data security concern is well-grounded in UK law. Under the UK General Data Protection Regulation (UK GDPR) and the Data Protection Act 2018, the Information Commissioner's Office (ICO) holds organisations accountable for how personal or commercially sensitive data is processed, including via third-party AI systems. Workers without a clear employer policy on approved tools are, in practice, being asked to make a legal judgement for which they have no training. The fact that 19% say training is inadequate and 12% do not know which tools are permitted is less a reflection of individual hesitation than of an organisational gap that employers are in a position to close.

For employees: making AI work in your favour

If you are among the non-managers who feel uncertain about feeding work documents into an AI tool, that caution has a rational basis. Data privacy is the concern raised most often in this survey, and the ICO's guidance on AI and data protection makes clear that responsibility for lawful processing sits with the organisation, not the individual worker. That said, there are practical steps you can take to navigate the current uncertainty.

• Find out what is approved before you start. Ask your line manager or IT team which AI tools your organisation has sanctioned and what categories of information can be entered. If no written policy exists, ask when one will be available. The ACAS (Advisory, Conciliation and Arbitration Service) code of practice on workplace technology policies, while primarily aimed at monitoring, establishes the principle that staff should know the rules governing digital tools.
• Request training and frame it as professional development. With 19% of respondents citing inadequate guidance, asking for structured training is a reasonable professional request, not a sign of difficulty. The Chartered Institute of Personnel and Development (CIPD) recommends that organisations treat AI upskilling as part of continuous professional development, not a one-off induction topic.
• Track the time you recover. If AI reliably frees up a meaningful number of hours per week, keep a record. That data strengthens your position in appraisals and salary conversations, particularly if your employer uses output-based performance frameworks.
• Review AI output before it leaves your desk. With 20% of workers concerned about accuracy, a quick check before submitting or sending anything AI-assisted protects your professional reputation and, in some sectors, your legal liability.

If you are considering a move and evaluating potential employers, how an organisation approaches AI governance is a reasonable question to raise at interview or in the research phase. Does the company have a published AI policy? Is there role-specific training? How does leadership communicate what efficiency gains will be used for? Reviews from current employees on verified platforms can show whether stated commitments reflect the daily reality for staff.

For HR: converting adoption into genuine productivity

For HR and reward professionals, the Adobe data reframes the AI conversation. The question is not whether staff are using AI tools; 68% already are. The question is whether reported time savings accumulate into measurable performance improvements, or whether they are offset by distrust, uneven adoption, and the friction that comes from absent policy.

The CIPD's annual Reward Management Survey consistently highlights that employee trust in organisational processes, including technology rollouts, correlates with retention and engagement outcomes. A 54-point gap between C-suite comfort and frontline comfort is the kind of structural disconnect that shows up later in pulse surveys and exit interview data if it goes unaddressed.

Organisations that handle AI adoption well tend to focus on three foundations:

• A clear, accessible data-handling policy aligned with ICO guidance. With privacy the top-cited barrier at 22%, staff need to know precisely which tools are approved and what information can go into them. Clear rules also reduce the pull towards unsanctioned tools, which carry greater organisational risk under UK GDPR and the Data Protection Act 2018.
• Targeted training, not generic sessions. The trust deficit is sharpest among non-managers and workers aged 55 and over. Role-specific training built around the tasks those groups actually perform, document drafting, spreadsheet analysis, scheduling, is more likely to shift confidence than a single company-wide briefing. ACAS guidance on managing change at work provides a useful framework for structuring these consultations, particularly in workplaces with recognised trade unions.
• Explicit communication on how saved time will be used. If employees suspect that efficiency gains will be absorbed into higher targets or used to justify headcount reductions, adoption slows and informal resistance grows. Stating clearly and in writing how recovered capacity will be directed, whether towards development time, higher-value work, or reduced unpaid overtime, is as much a retention decision as a technology one.

The gender gap also deserves specific attention in the HR review cycle. Where 45% of women report discomfort sharing sensitive documents with AI compared with 33% of men, monitoring adoption rates by team, role, and demographic characteristic is a prudent step. Uneven adoption that tracks along gender lines risks creating uneven recognition for productivity gains, which intersects with pay equity obligations under the Equality Act 2010 and, for organisations with 250 or more employees, gender pay gap reporting requirements.

The wider signal for UK HR strategy

A 68% adoption headline is easy to read as evidence that the transition to AI-assisted work is well under way. The fuller picture from this survey points elsewhere. The real division is between those who feel safe using AI in a professional context and those who do not, and it maps onto seniority, age, and gender in ways that most HR teams will recognise as pre-existing vulnerabilities in how change is communicated and supported.

For the Chartered Institute of Personnel and Development (CIPD) members designing workforce strategies for the next phase of AI integration, the 54-point trust gap between C-suite and frontline staff is the number that repays the most attention. Closing it is not primarily a technology problem. It is a policy, training, and communication challenge that falls squarely within the HR function's remit.

If you sit on the employee side, the practical move is to ask your employer what is approved, what training is available, and how any time savings will be recognised and redistributed. If you are on the HR side, the data suggests that investment in governance infrastructure and targeted upskilling is what converts a usage statistic into an actual productivity gain.