AI was meant to make work easier. For a lot of UK workers, it's doing the opposite. New research from YuLife and YouGov reveals a growing "AI productivity paradox": as companies pour investment into artificial intelligence, employees are reporting rising pressure, heavier workloads, and real uncertainty about where they fit into the future of work.The study, conducted among 1,184 UK workers, shows AI is beginning to reshape white-collar work, but not always in the way many expected.One in four workers (26%) who use AI tools in their workplace say AI is increasing pressure at work. Over one in five (23%) report that their workload has actually grown since AI tools were introduced. The technology promises major efficiency gains, but the data suggests the benefits aren't being felt equally across the workforce.White-collar workers feel the greatest riskOne of the most striking findings is a reversal of the usual automation story.Historically, automation has disrupted manual labour first. This time, it's professional roles in the firing line. 36% of ABC1 professionals who use AI tools at work say they worry AI could make their job redundant, compared with 25% of C2DE workers using AI. That's a 36% gap, and it points to a generational shift in which roles feel most exposed.Concern isn't universal, though. Among workers aged 55+ who use AI tools in their workplace, 73% say they aren't worried AI will make their role redundant within the next five years. And adoption among older workers is climbing: 22% of employees aged 55+ now use AI tools for writing and editing tasks like drafting emails or summarising reports.The productivity dividend gapThere's also a widening trust gap between employees and employers around how the benefits of AI will be shared.Despite heavy investment in AI across many organisations, 35% of professionals who use AI tools in their main workplace don't believe productivity gains from AI will be reinvested into their wellbeing or development. That's a third of AI users questioning whether the rewards of the technology will ever reach them."We are navigating a generational shift in how people work," says Tal Gilbert, CEO of YuLife. "The real question isn't whether AI will reshape white-collar roles because that is already happening. The challenge is whether we shape that transition deliberately, with people at the centre, or allow short-term efficiency economics to dictate the terms. Our research suggests too many employers have yet to strike that balance. At this scale of change, getting this right is too important to leave to chance."London: high adoption, high pressureIf you want to see where the rest of the UK is heading, look at London. The capital is acting as an early testing ground for AI-driven workplaces, and the data shows both the upside and the strain.AI adoption in London sits well above the national average. Only 35% of London workers say they don't use AI at work, compared with 50% nationally. But rapid integration comes with rising pressure:40% of London workers who use AI tools report increased performance pressure due to AI, compared with 26% nationally46% of London AI users are concerned about AI-driven redundancy in the next five years, compared with 34% nationally39% say their workload has increased since AI tools were introducedLondon isn't just adopting AI faster. It's feeling the squeeze faster too.Pressure, gender, and connection at workThe research also hints at how AI is shifting workplace dynamics more broadly.Men are significantly more likely than women to report increased performance pressure from AI, at 30% compared with 21%. And among Gen Z workers aged 18-28 who use AI, 36% say AI is making workplace interactions feel more transactional. That raises real questions about what happens to collaboration, trust, and connection when more of the work runs through a machine.What this means for employersThe picture the research paints isn't anti-AI. Adoption is happening, and it's happening fast. But the data makes clear that the way organisations introduce AI matters just as much as whether they introduce it. Workers are telling us they feel the pressure, they doubt they'll share in the gains, and in some cases they're questioning their place in the future of their own industries.Employers who want to get this right have to treat the human side of AI transition as seriously as the technology itself. That means being deliberate about how productivity gains are reinvested, honest about what AI means for roles, and thoughtful about the culture and connection that make work worth doing in the first place.At this scale of change, the cost of getting it wrong is too high to ignore.Research conducted by YouGov Plc. Total sample size was 2,088 adults, of which 1,184 are working adults. Fieldwork was undertaken between 2nd-3rd March 2026. The survey was carried out online. Figures have been weighted and are representative of all UK adults (aged 18+).