51% of Scottish Workers Don't Trust GenAI, And It's Costing Scotland a Productivity Gap With London

Deloitte's first GenAI Workforce Survey, drawn from 25,000 UK workers, confirms most Scottish workers are now using generative AI on the job. But half still don't trust it, and that hesitation is opening a measurable competitiveness gap between Scotland and the capital that Scottish SMEs cannot afford to ignore.

The headline number from Deloitte's inaugural GenAI Workforce Survey is this: 51% of Scottish generative AI users cite trust as a significant barrier. That is not a fringe concern from technophobes. That is the majority of people already using the tools, sitting at their desks, running the outputs through a filter of scepticism that slows everything down. The survey, conducted between 7 and 10 June 2026 across 25,000 workers in the UK, is the most detailed picture yet of how GenAI is actually landing in British workplaces, not how vendors wish it were landing.

The productivity gap with London is the part that should land hardest for anyone running a business north of the border. Scottish workers are using GenAI, but the survey points to London-based workers extracting significantly more output from the same tools. The reasons are structural: London organisations are investing more heavily in training, internal AI champions, and clear governance frameworks that tell employees what the tools are for and, critically, what they are allowed to do with them. Without that scaffolding, workers default to caution. Caution compounds. The gap widens.

This matters beyond individual firms. According to the Scottish Government's AI Strategy, Scotland has committed to becoming a leading AI nation, with digital skills and AI adoption cited as central to closing the productivity gap that has long separated the Scottish economy from the UK's south-east. The Deloitte data suggests that ambition and reality are not yet aligned. Scottish Enterprise and Skills Development Scotland have both flagged AI upskilling as a priority, but survey evidence of this scale makes it harder to treat the gap as a minor lag rather than a structural problem.

The trust issue is not irrational, and it is worth treating it seriously rather than dismissing it as technophobia. Workers worry about accuracy, about data privacy, about professional liability, particularly in regulated sectors like healthcare, financial services, and education, all of which are significant employers across Scotland. Research from the Alan Turing Institute has consistently shown that trust in AI systems is closely tied to transparency: workers trust tools they understand, and distrust black boxes they have been handed without explanation. The fix is not a motivational poster about embracing change. It is clear internal policy, visible leadership, and training that is practical rather than theoretical.

For Scottish SMEs, the read-across is direct. If large organisations with dedicated IT, HR, and L&D teams are struggling to turn GenAI adoption into productivity gains, smaller businesses operating without those resources face an even steeper climb, unless they move deliberately. The opportunity here is real. A ten-person firm that builds a culture of confident, trusted AI use today does not just catch up with London; it competes with companies ten times its size. The tools are available, most are affordable, and several are free. The barrier is not access. It is internal clarity about why, how, and when to use them, and that is entirely within an SME owner's control to fix.

What this means for you: If your team is using GenAI tools but quietly second-guessing every output, you have a trust problem, not a technology problem. Set a clear, written internal policy this week: what tools are approved, what data stays out of them, and what outputs need a human check before they go anywhere. That one document will do more for your productivity than any new subscription. Then find one person in your team to own it, and make them your internal AI lead. You do not need a budget to do this. You need an afternoon.