London workers to be hardest hit by AI
Key Points
- The OECD Employment Outlook 2026 estimates that between 16% and more than 70% of jobs are highly exposed to generative AI depending on the region.
- ICT and finance-heavy urban areas are set to be most affected.
- The OECD warns this could widen urban and rural divides.
- Wage inequality across UK local labour markets has more than tripled since the 1970s.
- AI could add 0.25 to 0.6 percentage points to annual productivity growth over the next decade.
Generative AI could deepen the divide between big cities and the rest of the country, the OECD has warned, as new figures show enormous regional differences in how exposed jobs are to the technology.
In its Employment Outlook 2026, published on 3 July, the OECD estimates that the share of workers whose jobs are highly exposed to generative AI ranges from around 16% in some regions to more than 70% in others.
The gap reflects differences in local industrial makeup. Regions with high concentrations of employment in ICT, finance and education, and in occupations built around cognitive and non-routine tasks, face the highest exposure.
In practice, that means capital cities and major urban centres, which already offer the strongest job prospects, are also where AI’s effects will land first and hardest.
The OECD warned that these regional differences in AI exposure could exacerbate existing urban and rural divides in incomes and labour market outcomes.
A divide that is already growing
The warning comes against a backdrop of widening regional inequality in the UK.
Research cited in the report shows that wage inequality across local labour markets has more than tripled in the UK since the 1970s, a sharper rise than in Canada, Germany, or the United States, where it roughly doubled.
The OECD also found that employment rates between regions differ by more than 20 percentage points in over half of OECD countries, and that people in the poorest regions are nearly 40% more likely to remain in the lowest income group over five years than those in the most affluent areas.
Job destroyer or productivity boost
Evidence on what AI exposure actually means for workers is still emerging, and the report notes it cuts both ways.
US data cited by the researchers found that employment among younger, early-career workers in the occupations most exposed to generative AI has declined relative to less exposed roles.
Other research cited by the OECD found that AI adoption has not yet caused widespread job losses, and is instead associated with productivity gains and employment growth in high-exposure occupations.
OECD estimates suggest AI could add between 0.25 and 0.6 percentage points to annual productivity growth over the next decade.
Whether those gains flow to exposed regions or drain away from everywhere else will depend heavily on policy.
The OECD pointed to place-based policies, such as Canada’s programmes supporting AI development and diffusion in specific regions, as one way governments can spread the technology’s benefits beyond the biggest cities.
