AI Use Cuts Hiring, Not Layoffs, Says Bank of England

The Bank of England published on November 5, in the Monetary Policy Report, the most direct reading by a major central bank on AI and employment. The impact of technology on total employment in the UK is currently close to zero. However, the bank's regional network of agents has heard from companies that adopting AI reduces the likelihood of hiring or replacing departing employees, without significantly increasing the likelihood of layoffs.

The Cut That Doesn’t Show in the Count

The public debate measures AI by layoff announcements. On October 22, Barclays announced it would eliminate about 5,000 positions this year, nearly 5% of its workforce, and CEO C.S. Venkatakrishnan attributed "most of the efficiency gains" to AI. This is the figure that makes headlines. What the Bank of England describes is different: the vacancy that remains unfilled when someone resigns and is never replaced, or the shrinking trainee program without notice.

This difference changes who bears the cost. A layoff affects those who are employed, appears in layoff trackers, and incurs severance costs. A hiring that doesn't happen impacts those who haven’t yet entered the workforce, particularly young individuals starting their careers in administrative and analytical roles, and leaves no trace in any layoff statistics. Aggregate employment can remain stable for years while the entry door narrows.

The Opposing Argument, by Name

The thesis that AI is already destroying jobs at scale has supporters with data. A Morgan Stanley survey released in January, involving companies that had used AI for at least a year, found a net loss of 8% of jobs in the UK over 12 months, double the international average and the worst result among the five economies studied. Rachel Fletcher, one of the authors, called the finding a "early warning sign." Treating this as alarmism would be a mistake.

Blind skepticism is also erroneous. Companies in the Bank of England's Decision Maker Panel estimate that AI has increased their productivity by about 0.5% over three years, a small gain insufficient to justify a workforce restructuring. The bank attributes most of the recent improvement in British productivity to other factors, including businesses unwinding post-pandemic hiring excesses and responding to rising employment costs. Some effects attributed to AI may actually be driven by payroll costs. This point weakens the thesis that AI is cannibalizing office work and needs to be stated.

Two Markets, Opposite Signals

In the United States, October data indicates the opposite direction. The employment report released on November 6 showed that the professional and business services sector, which includes consulting, accounting, and legal services, created 70,000 jobs in the month, compared to an average of 15,000 per month over the previous 12 months. The American economy added 165,000 jobs, above projections, and unemployment fell to 4.1%. These are preliminary numbers, subject to revision.

The same Morgan Stanley survey had previously shown this divergence in January: the US was the only one of the five countries where AI-using companies reported a net job gain, with productivity gains similar to those in the UK. The same technology, yet opposite results in the job market.

In the UK, the channel is cost. With payroll more expensive, AI provides British managers a concrete reason not to replace those who leave, and Barclays' cuts are concentrated in the country. In the US, with strong demand for professional services, the same tool is used to serve more clients with the existing team, and hiring continues.

The Axis of Comparison

The question "Is AI destroying jobs?" tends to produce self-cancelling answers, as it measures the stock of employees, which has hardly moved. The question that the data from London and Washington better answer is another: how much of the inflow each company decided not to open, and why.

For those leading a consulting firm or a bank, this indicator exists but rarely reaches the board: the replacement rate of those leaving, by role and level. It shows, before any layoff announcement, where AI, or payroll costs disguised as AI, is actually replacing people. The Bank of England has begun tracking exactly this movement, and companies that measure it internally will know first which of the two explanations applies to their own workforce.