Coface: AI Threatens Britain's High Salaries and Tax Revenue

A study by Coface and the Observatoire des Emplois Menacés et Émergents found that the UK is more exposed than other developed countries to AI task automation. The risk is not limited to employment: if the pressure falls on high-paid office professions, the budget could lose part of its key tax revenue.

Why Britain is vulnerable

According to the study, among 12 developed economies the UK has the highest share of tasks exposed to automation: about 20% of the total workload. For comparison, France came in closer to 16%, while Germany and the US were at roughly 17%.

The authors stress one important point: this is not a direct forecast of instant unemployment but a measure of the technical exposure of tasks to automation. Still, for a country with a large share of financial, legal, managerial and information services, that is enough to speak of a systemic risk.

Another problem is that AI hits not the cheapest roles but those where money, taxes and corporate expertise are concentrated. The report says that in the professions held by the top-earning 10% of workers, 20–25% of task content is at risk. This is especially sensitive for the British economy: London and other business centers have grown for years as hubs for headquarters, financial services, consulting, IT and media. If AI systems take over part of this work, companies and infrastructure owners will capture the productivity gains first, while the payroll tax base may shrink.

Where the blow lands hardest

AI is best suited to tasks where information is turned into information: data analysis, drafting text, structuring documents, calculations, checking standard scenarios. That is why the study finds the highest exposure in engineering and computing professions at 29%, legal and financial at 27%, creative and content roles at 27%, and managerial and administrative at 24%.

This matches the structure of the British economy, where service industries are already adopting AI more actively than manufacturing and construction. The government's own assessment puts the skew plainly: AI is more likely to affect professions involving cognitive rather than physical tasks. So a builder, driver or care worker is for now better protected than an analyst, lawyer, editor or office manager. For Britain this is an especially unpleasant scenario, because the country has historically relied on expensive intellectual labor and on its role as Europe's hub for international business.

• Financial and related professional services contributed £110.2 billion in taxes to the UK budget in 2024, or 12.3% of all receipts.
• In the coming years, 13% of professions may cross the threshold at which more than 30% of their tasks become automated.
• According to British authorities, current AI technologies already potentially affect 10–30% of jobs, with skilled cognitive roles hit harder.
• By mid-2025, vacancies in professions with high AI exposure were about 5.5% below the expected level.

So far this does not look like a one-off wave of layoffs. Rather, the market is sending early signals: companies are hiring more cautiously for routine intellectual positions while complaining about a shortage of new skills. In the 2025 DSIT survey, 97% of employers reported at least one AI skills shortage, and 57% reported a lack of technical skills. This means some tasks will be automated faster than the education system and the labor market can rebuild career paths.

What this means

For the UK, the AI debate is becoming not just a conversation about productivity growth but a question of how to fund the state in an economy where less income is tied to human labor. If high-paid office roles shrink faster than new ones appear, the country will have to retrain people, support transitions between professions and look for new sources of tax at the same time, closer to capital, profit and AI infrastructure.