OP-ED - When "lower value human capital" meets the AI bill that costs more than people
Standard Chartered CEO Bill Winters apologised for the phrase, not the thinking behind it. Meanwhile, the numbers from Microsoft, Uber and Nvidia suggest humans may have been the bargain all along.
In other news, last week's proffer that the cheaper option (to AI deployment) might be human just found an unlikely foil.
Andile Masuku
Standard Chartered chief executive Bill Winters has apologised for describing a Hong Kong banking migration as replacing, in some cases, "lower value human capital" with the financial and investment capital being put behind AI innovation, among other things.
The backlash to his initial statements was immediate and global. Former Singapore President Halimah Yacob called it "disturbing" and "demeaning". LinkedIn lit up. By last Friday, Winters had apologised acknowledging that his "choice of words" had "caused upset to some colleagues". Next time, he'll no doubt stick to approved PR terminology like rightsizing, reskilling, redeployment, transition packages, clear notice, etc.
Winters' backtrack notwithstanding, does his slip of the lip betray hopelessly entrenched boardroom sensibilities towards the AI opportunity?
At the minute, markets are clearly bullish on AI deployments enabling workforce cost reduction in the short to medium term. Winters et al are playing to the shareholders (read the money).
Awkwardly, I did much the same thing in my own column last week. I basically priced out the human against the agentic system. I wrote about "a senior analyst, a careful sub-editor, a junior compliance officer, a developer who reads the codebase" and stacked their "fully-loaded cost" against "five-figure monthly token bills." Now granted, I knew better than to go with "lower value human capital".
How effortlessly the language of valuation colonises the way we see people, even for someone writing critically about the trend. Winters didn't invent "lower value human capital", he just forgot to use the approved corporate translation.
Meanwhile, the market waits for no man, whatever their rhetoric.
Recent reports of Microsoft cancelling most internal Claude Code licences by June 2026, shifting engineers to its own GitHub Copilot CLI due to high token-based costs at scale, alongside similar budget overruns at Uber where AI spending exhausted the entire 2026 budget allocation by April—Uber's CTO noting heavy usage (up to USD 2,000 monthly per power user). And an Nvidia VP stating AI compute costs now exceed employee expenses for his team.
Some analysts reckon this reflects a transitional phase where token pricing is driving enterprises towards internal tools and cost controls, rather than outright rejection of AI, as companies balance adoption with budgeting realities amid forecasts of rising total spend.
Note to self: beware the pragmatic utility of gauging "how AI is going" by how convincingly corporates are converting its adoption to business gains, i.e. bottom-line impacting efficiencies or business growth amplifiers. Even as professionals the world over are trying to establish how to leverage the technology to advance their aspirations, Western graduates protest what appears to be their promised career disruption, and while philosophers and policymakers weigh whether the tech is a net positive for society and what safeguards should be in place to prevent institutional greed, selfishness and unbridled opportunism from swallowing us all whole.
Incidentally, I write this on Africa Day, which happens to fall on the UK's Spring Bank Holiday. Maybe this is simply a philosophical waxing that duly ends with King Solomon's proverb:
What is crooked cannot be made straight, and what is lacking cannot be counted.
In this case, the crooked thing is the shared language that makes it normal to call any human "lower value", or to discreetly price them against a token bill and call it progress.
And what's lacking?
Perhaps the humility to admit that even those of us critiquing Winters' slip are still standing inside the same capitalist calculator. The money doesn't care who apologises.
Editorial Note: A version of this opinion editorial was first published by Business Report on 26 May 2026.