City Hall’s latest assessment of AI exposure in the capital has put more than one million London roles in the high-risk category, or just over a fifth of the workforce. Narwhal Labs chief executive Luke Sartain says the figure understates the scale of disruption already under way, arguing that AI exposure now extends far beyond the headline estimate and is already reshaping routine knowledge work.
The Bristol company said the focus should not be limited to jobs formally classed as “at risk”, but to roles already being compressed by automation across structured, repeatable, digital tasks. Sartain cites IMF data suggesting around 60% of jobs in advanced economies are exposed to AI, with the UK among the most impacted markets. Sartain said London’s one-in-five estimate should be seen as a conservative measure rather than a full account of the labour market shift now taking place.
He said: “The ‘one in five’ figure is already out of date. For most knowledge work, AI displacement isn’t coming — it’s already here. Roles built around coordination, communication and routine decision-making are being compressed first. When work is structured, repeatable and digital, AI can now do it faster and continuously.”
Sartain said the impact is already being felt in jobs that sit between systems and customers, including administrative, support, and coordination-heavy functions. He also said the next phase of disruption will move beyond software-based work. As embodied AI and humanoid systems enter physical environments, he said, the number of roles affected could increase sharply in some categories, with multiple jobs eventually replaced by a single system.
Sartain said companies that treat AI as core infrastructure will move faster, operate more leanly, and capture more value, while those that continue to treat it as a side experiment risk falling behind. City Hall’s figures offer one benchmark for London, but Narwhal Labs argues the shift is already broader and moving faster than many organisations have accounted for.
You must be logged in to post a comment.