AI at work isn't a revolution. It's a vacuum.

More than four in five UK workers now use AI at work. Fewer than one in twenty use it to change how they actually work.

That second number matters more.

We mistook adoption for competence

We’ve decided the AI story at work is one of adoption. But most of that use is searching and summarising, the shallow end. Those figures come from EY’s survey of UK workers, which also found only about one in ten getting proper training in it. In the government’s own survey, only about one in five felt confident using AI at work. And at the end of 2025, only a quarter of UK businesses told the ONS they used AI at all, while most of their staff already were.

So what’s happening? People are teaching themselves, quietly. Microsoft’s UK research found that 71% of employees have used AI tools their employer never approved, and half of all employees still do it every week. They’re not being reckless. Nobody gave them an alternative.

There’s a cost to working like that. When the tool you lean on is one you’re not sure you’re allowed to use, you carry a low-level worry alongside the work: am I going to be caught, blamed, or left exposed if this goes wrong? It’s a small worry, but it’s there every day, and it makes the work harder.

This isn’t an AI success story. It’s a vacuum.

Responsibility without support is how you create stress

We’ve handed people one more thing to be quietly responsible for, with no support and no permission. Then we act surprised when it goes wrong, and blame them when it does.

Mark Fisher had a name for this. He called it the privatisation of stress: the way our economy takes problems that are collective and structural and hands them back to us as private burdens to manage alone. If you feel overwhelmed, that’s treated as yours to fix, not something the system has to answer for.

AI at work is the privatisation of stress happening right now. Workers are left to deal with it themselves, and the organisation does nothing. It’s the same pattern behind so much burnout, where the person is handed a list of things to cope with and the conditions that caused the strain go untouched.

The skills gap in the headlines isn’t really in your people. It’s in everything around them: the strategy, the rules, and the basic decision to say “here’s how we do this, and here’s the help”.

Have you given them permission, or just the tools?

So before the next headline about AI-ready workforces: have you given anyone the training, the tools and the permission to use this well? Or have you left them to it, and called that progress?