From AI adoption to supercharged progress: the skills challenge facing Britain

Prime Minister Keir Starmer’s London Tech Week keynote was, at its heart, a speech about choice. Not just whether Britain embraces AI, but how it chooses to do it. Do we ignore the revolution and hope it passes us by? Do we remove the guardrails and accept the consequences, no matter who gets left behind? Or do we back British innovators, build sovereign capability, protect people through change, and make sure the benefits are felt far beyond London?

That third path is the right ambition. In fact, I hope this speech becomes a turning point for British businesses because the future the Prime Minister described points towards a more expansive possibility: not simply an AI-enabled economy, but supercharged progress.

Infrastructure is essential, but it’s not enough

The PM rightly focused on national conditions: capital, compute, procurement, research and development, regulation, safety, and trade. These are essential. Britain can’t lead in AI if promising companies start here, scale elsewhere, and sell elsewhere. Sovereign compute capability, including the government’s commitment to purchase specialist AI chips, is a serious signal that the UK wants to build the foundations of the future, not rent them indefinitely from others.

But infrastructure alone won’t create the future we want. The decisive question now is whether our workforce can move as quickly as the technology.

In my view, there are not only three options in front of Britain. There are also three possible future scenarios for businesses and workers.

Scenario one: supercharged progress

This is the most exciting scenario. New occupations emerge and scale quickly. Humans don’t just ‘use AI’ in a narrow sense. They direct portfolios of AI agents and machines. They become orchestrators, designers, challengers, and decision-makers in increasingly complex human-machine systems. This is where productivity gains become meaningful, where innovation accelerates, and where entirely new value chains are created.

But this doesn’t happen by learning a handful of prompts. It requires targeted continuous learning. It requires people to understand how to question AI, how to test outputs, how to recognise weak reasoning, and how to combine domain expertise with technical confidence.

The next wave of AI skills is not about creating more vibe coders who need someone else to check their work. It’s about creating true citizen developers: people with enough technical understanding to build, interrogate, and improve AI-enabled workflows responsibly.

Technical training is no longer just for technical audiences. AI is democratising access to technical capability, but democratisation without education is risky. If everyone can build, everyone also needs to understand enough to build well.

Scenario two: the co-pilot economy

I don’t mean Microsoft Copilot, but the broader model of people working in human-AI pairings to improve productivity.

The co-pilot economy can create real gains. It can reduce repetitive work, improve speed, and help people access knowledge faster. But the question is whether incremental productivity is enough to sustain competitive advantage.

If British businesses stop here while others move towards agent orchestration and AI-native operating models, we risk mistaking early adoption for transformation.

Scenario three: stalled progress and worker displacement

This will happen if the workforce is lacking critical skills for either of the scenarios above.

Worker displacement is what happens when businesses race to automate as a stopgap, but education and reskilling systems can’t keep pace. Workers are displaced faster than new pathways are created and while productivity may rise in pockets, trust falls everywhere.

Workforce learning needs to be a priority if businesses want to avoid this. Otherwise, AI adoption advances, investment is available, leaders are ambitious, yet organisations can’t move forward because their people lack the critical skills to implement change.

Workforce learning must become a strategic priority

The government’s target to upskill 7.5 million workers with AI training by 2030, and the reported progress of 1.7 million already trained, is encouraging. But businesses can’t wait for national programmes alone.

Every organisation needs to ask: which AI capabilities do our people need now, what will they need next, and how quickly can we build them?

The answer can’t be generic awareness training. It must be role-specific, continuous, and practical. People need to learn how AI changes the work they actually do. They need to understand where automation helps, where human judgement matters, where risks sit, and how to work confidently with tools that will keep evolving.

The real test for Britain

Britain has many of the ingredients it needs: talent, research, capital, ambition, and a growing recognition that AI is an industrial strategy issue, not just a technology issue. The choice now is whether we connect those ingredients to a national learning agenda at the scale required.

If we do, we can move beyond a co-pilot economy into supercharged progress. If we don’t, we risk stalled transformation or displacement by default.