Liz Kendall doesn't use AI at work. The Work and Pensions Secretary admitted this recently, and it’s a bit of a head-scratcher. You’d think the person overseeing a massive department responsible for modernizing the UK's welfare state would be experiment Number One for these tools. Instead, we have a clear divide between the high-level rhetoric of "AI-powered public services" and the actual daily habits of the people in charge. It’s not just a Kendall thing. It points to a much bigger problem in how the public sector views technology.
If you’re wondering why this matters, look at the budget. The government talks about billions in efficiency savings. They want to automate back-office tasks and speed up benefit processing. But if the Secretary of State isn't even using a basic LLM to summarize a long policy briefing or draft a memo, how can she effectively lead a digital transformation? It feels like someone trying to teach a driving lesson while refusing to touch the steering wheel.
The gap between policy and practice
Kendall’s admission happened during an interview where she was asked about her personal tech use. She was honest. She doesn't use it. While honesty is refreshing in politics, this specific brand of honesty highlights a massive cultural lag. The Department for Work and Pensions (DWP) is a behemoth. It handles millions of claims. It’s exactly the kind of place where AI could actually do some good, yet the leadership remains disconnected from the tech itself.
We see this everywhere in government. Ministers announce "AI hubs" and "innovation units" while their own offices still run on paper and static PDFs. You can't lead a tech revolution from the sidelines. To understand the risks—hallucinations, bias, data privacy—you need to actually get your hands dirty with the tools. Otherwise, you're just reading scripts written by consultants.
Why the DWP is the perfect AI testing ground
The DWP is basically a giant data processing machine. People apply for money. The department checks if they’re eligible. They pay them. It sounds simple, but the bureaucracy is legendary for being slow and opaque.
Think about what happens when a claimant calls a helpline. They're often stuck in queues for hours. An AI assistant could handle the basic stuff—checking payment dates or updating addresses—leaving humans to deal with the complex, emotional cases. But to implement that, the leadership needs to understand what "good" AI looks like. Kendall's distance from the technology suggests she might not know the difference between a useful automation and a total disaster.
There’s also the issue of fraud detection. The DWP already uses algorithms to flag suspicious claims. This has already caused controversy. Without a deep, personal understanding of how these systems work, ministers are ill-equipped to defend them or, more importantly, to know when to shut them down because they're being unfair.
The fear of the machine
Why aren't they using it? It’s likely a mix of security concerns and a fear of looking "lazy." There’s a persistent myth that using AI for writing or analysis is a shortcut for people who don't want to do the work. That’s nonsense. In 2026, using AI is about efficiency. It’s about getting through the noise so you can focus on the hard decisions.
In the private sector, a CEO who refused to use modern communication or analytical tools would be seen as a dinosaur. In Westminster, it’s just another Tuesday. This tech-aversion trickles down. If the boss doesn't care, why should the civil servants? It creates a culture where "innovation" is just a buzzword for the annual report rather than a daily practice.
Breaking the digital ceiling
Government departments need more than just a "Minister for AI." They need every minister to be tech-literate. We aren't talking about coding in Python. We're talking about understanding the utility and the limitations of the software that is currently reshaping the global economy.
When Kendall says she doesn't use it, she's signaling that it isn't a priority for her. That's a mistake. The DWP is currently facing massive pressure to cut costs while improving service delivery. AI is the only way to do both simultaneously.
What happens when the leadership stays analog
When leaders stay analog, policy suffers. You get regulations that don't match reality. You get procurement processes that buy overhyped, expensive junk because nobody in the room knows how to ask the right technical questions.
Look at the Post Office Horizon scandal. While that wasn't "AI" in the modern sense, it was a failure of leadership to understand a technical system. They trusted the "black box" because they didn't know how to look inside it. Kendall’s refusal to engage with AI at a personal level makes her vulnerable to the same kind of blind trust—or blind skepticism. Neither is good for the public.
Real-world AI use for a minister
If Kendall wanted to start tomorrow, here is what it would look like. She could use a secure, private instance of a model to:
- Summarize 200-page stakeholder reports into five key bullet points.
- Draft initial responses to constituent letters, which she then edits for tone.
- Run "what-if" scenarios on policy changes to see potential impacts on different demographics.
- Check her own speeches for repetitive language or clarity.
None of this is "replacing" her. It’s augmenting her. It’s giving her more time to actually meet with the people the DWP is supposed to help.
The risk of being left behind
The UK wants to be a global leader in AI. We host summits. We invite Elon Musk to Bletchley Park. We talk a big game about safety and ethics. But if our own Cabinet ministers aren't part of the user base, the rest of the world will stop taking us seriously.
You can't regulate what you don't use. You can't inspire a workforce to modernize if you're stuck in 2010. Kendall’s stance isn't just a personal choice; it’s a policy statement. It says "this isn't for us." And that is exactly the wrong message to send right now.
Taking the first step
If you find yourself in a leadership position and you're "pulling a Kendall," it's time to stop. The world is moving too fast for the "I'm not a tech person" excuse to hold weight anymore.
Start small. Find a repetitive task you hate. See if a tool can help. You don't need a departmental strategy to start using a tool for your own productivity.
Leaders need to:
- Demand a "sandbox" environment where they can test tools without risking public data.
- Stop treating IT as a support function and start treating it as a core part of policy.
- Admit they don't know everything and bring in people who do—not just as advisors, but as decision-makers.
The DWP has a long way to go to fix its reputation. Liz Kendall has a chance to be the one who finally drags it into the modern era. But she can't do that if she stays on the outside looking in. It’s time to log on.
