The people best prepared for the AI age learned the hard way.

Disabled Entrepreneur

AI is threatening the jobs that disabled people depend on most. But their lifetime of navigating a world not built for them might be exactly the skill set the rest of us now need.

Nearly a million young people in the UK are currently not in work, education or training. The highest figure in a decade, and still rising.

At the same time, entry-level job postings are running 45% below their five-year average. Graduate roles in banking and finance have fallen 75% since 2019. Software development roles down 65%. Accounting down 54%. PwC cut 200 entry-level positions in late 2025 and publicly said AI was the reason, noting that generative AI can now draft reports and review documents that used to be given to interns and trainees.

The first rung on the career ladder is disappearing. And it’s disappearing fast.

Now, picture a teenager in the middle of all this. Smart. Curious. Full of potential.

But school isn’t really working for them either. Not because they can’t do the work. Because the system wasn’t designed with them in mind. They’re a disabled young person, waiting — probably for two years or more — for an autism or ADHD assessment that might finally explain what’s been going on. While they wait, they fall further behind. Children on SEN support are nearly two years behind their peers by the time GCSEs come around. Those with more complex needs are more than 3 years behind. Some get suspended before anyone has worked out what support they actually need.

The government’s Schools White Paper, published last month, does at least signal that change is coming. Every child with SEND will gain a new legal right to an Individual Support Plan. There’s an additional funding commitment of £7 billion by 2028-29. Disability charities have cautiously welcomed the direction of travel. But the new system doesn’t begin until September 2029. For a teenager waiting for an assessment today, that timeline is cold comfort.

The education system has one measuring stick. And for a lot of disabled young people, it’s still the wrong one.

Some of them make it to university. And here, genuinely, things start to improve. Reasonable adjustments exist in law — extra time in exams, accessible formats, and assistive technology. The attainment gap between disabled and non-disabled students narrows considerably compared to school. Progress, of a kind.

The catch? Only 39% of disabled students say they actually receive all the adjustments they’re entitled to. Three in five don’t get what they’ve been promised. The system tries harder at university. It just doesn’t always follow through.

Still. Some of those young people graduate and find their footing — often in office-based work. Desk-based. Remote-friendly. Roles where what matters is what you can do, not whether you can manage four flights of stairs or a packed commute. Since COVID, remote and hybrid working has opened doors that were previously shut. For many disabled people, it’s been the closest thing to a level playing field they’ve ever experienced at work.

And now AI is coming for exactly those jobs.

Data processing. Admin. Document drafting. Customer service. Basic analysis. These are the roles most immediately in AI’s sights — and they’re disproportionately the ones disabled people have been able to access. Meanwhile, the work AI genuinely can’t replicate — your plumber, your electrician, your joiner working in a tight Victorian terrace — often requires physical capability that simply isn’t possible for a significant number of disabled people.

The accessible jobs are at risk. The resilient ones remain out of reach. It’s a genuinely uncomfortable picture, and one that isn’t getting nearly enough attention.

But here’s something I think gets missed in most of these conversations.

Some disabled people have already found their own answer. They’ve stopped waiting for employers to accommodate them and started building something themselves. Disabled workers are already slightly more likely to be self-employed than non-disabled workers. And when you think about it, that makes complete sense. Self-employment lets you design work around your needs — your hours, your environment, your energy. No waiting for an employer to make reasonable adjustments, because you are the employer.

A major investigation — the Lilac Review, published in 2024 — found that the barriers facing disabled entrepreneurs cost the UK economy an estimated £230 billion in lost revenue. That’s not a niche statistic. That’s a serious economic failure hiding in plain sight. The review surveyed 750 disabled entrepreneurs and found that 73% want to grow their businesses. The ambition is very much there. The system just isn’t supporting it.

Some support is emerging. Access to work can fund equipment and adaptations for the self-employed. Hatch Enterprise runs accessible entrepreneurship programmes for disabled people wanting to start a business. A Disability Finance Code for Entrepreneurship launched in December 2024, developed with the Department for Business and Trade and the British Business Bank. And the LILAC Centre — the UK’s first business incubator dedicated specifically to disabled entrepreneurs — has just opened its doors.

These are encouraging developments. But they’re recent, not yet at scale, and not joined up. The Lilac Review found that 84% of disabled founders feel they don’t have equal access to the same opportunities as non-disabled founders. And there is still no clear, accessible pathway for someone who becomes disabled in their forties, or who loses a job they depended on, that simply says: here is how you start again. That pathway needs to be built.

This is where AI re-enters the story — but this time as an opportunity rather than a threat. The tools disrupting office work are the same ones that can dramatically lower the barriers to running your own business. AI can handle your marketing copy, your social media, your admin, your invoicing. For a disabled entrepreneur managing limited time or energy, that’s not a small thing. It means parts of running a business that once required significant capital or a whole team can now be handled by one person with a laptop and the right tools.

Which brings me back to that teenager.

Think about what they’ve actually learned on their journey. They’ve navigated a system that measured them by the wrong ruler. Waited years for support that arrived late and often incompletely. Found ways through, over and around obstacles that were never supposed to be there. That’s not a sob story. That’s a skill set. Adaptability. Stickability. Creative problem-solving when the obvious route is blocked. The ability to keep going when the process wasn’t designed for you.

Those are exactly the qualities the AI age will demand of everyone.

Which makes it all the more important that we rethink what we’re teaching. Not just for disabled young people, but for all young people. The current model — steer everyone towards a degree in a specific subject and hope the job market holds — is looking increasingly shaky. AI is already transforming law, programming, accountancy, financial analysis, and junior medical roles. The subjects we’ve been pointing young people towards for decades are changing faster than the curriculum can keep up with.

What if instead we taught AI literacy alongside traditional subjects? What if entrepreneurship, marketing, social media and business development were treated as serious skills, not optional extras? What if we helped young disabled people — who may never find the traditional employment route straightforward — think about building something of their own as a genuine, supported option rather than a last resort?

There are signs of movement. A new computing GCSE with AI literacy built in. Government investment in digital skills is reaching into schools. But disabled students are already being flagged as one of the groups most at risk of being left behind in this transition. A new technology gap, layered on top of an existing attainment gap, would be a serious mistake.

Disabled people didn’t choose to develop resilience and adaptability. The system made it necessary.

But in a world that’s going to demand those qualities from all of us, perhaps it’s finally time we designed systems that support disabled people properly — not just because it’s the decent thing to do, though it is — but because we genuinely can’t afford not to.