Generated by Codex with GPT-5

The managerial problem behind AI anxiety

The article treats a badly phrased corporate comment as a useful case study in a broader management problem. Standard Chartered’s chief executive, Bill Winters, drew criticism after describing automation as replacing “lower-value human capital” with financial capital. The backlash was predictable, not because the bank was wrong to talk about automation, but because the language made people sound like assets to be sorted into valuable and disposable categories.

That kind of wording lands badly in the generative-AI era. Bosses want employees to adopt AI because competitors are adopting it, productivity pressure is real, and many repetitive office tasks can now be automated. Employees can see the potential benefits too, but they also see the threat to their jobs. The article notes that an AI backlash is already visible in workplace culture, student reactions, and online jokes about becoming obsolete. The acronym FOBO, or fear of becoming obsolete, captures the mood: people are not merely being asked to learn a new tool; they are being asked to embrace a technology that may reduce their own bargaining power.

The central point is that managers cannot talk about AI as if it were only a capital-allocation problem. It is also a trust problem. When leaders sound indifferent to the human consequences, they make adoption harder, because workers have every reason to doubt that the gains will be shared or that displaced people will be treated seriously.

Why trust matters more than slogans

The article draws on research about job insecurity to explain why tone and credibility matter. Job insecurity is bad for workers: it is associated with worse health, lower satisfaction, and greater stress. But it is also bad for firms. Although executives may imagine that fear will make employees work harder, the evidence points the other way. Insecure employees tend to perform worse, not better.

Trust can soften that damage. Research based on a large British workplace survey found that when workers saw managers as honest, reliable, and fair, the negative effect of insecurity on commitment was weaker. That does not make disruption painless. It means employees are more likely to stay engaged when they believe management is being straight with them and has a plausible plan.

That is where Standard Chartered’s example becomes more mixed. The phrase that caused trouble was clumsy and revealing. “Human capital” is already a cold way to describe colleagues; attaching “lower-value” to it makes the problem worse. It also misstates how AI exposure works. The jobs most vulnerable to automation are not always low-status or low-skill. Software engineers, analysts, customer-service workers, back-office staff, and other white-collar employees can all be exposed if their tasks are repetitive, easy to check, or increasingly reproducible by models.

But the fuller version of Winters’s comments included something more defensible: the bank had tried to identify new roles for employees affected by automation who wanted to remain at the firm. That matters. The article argues that credible preparation is more useful than bland reassurance. Employees do not need to be told that everything will be fine. They need to know which human skills will still matter, what work is likely to change, and how the company will help people move before their old roles disappear.

Preparing people for the work that remains

The article’s practical recommendation is that firms should combine honesty about AI disruption with active preparation. That means identifying jobs likely to be reshaped by automation, mapping the skills that will remain valuable, and helping employees acquire them early enough to make a difference.

Some roles can be made less vulnerable by moving people toward more interpersonal, judgment-heavy, or revenue-generating work. The article points to DBS, a Singaporean bank, which has helped workers shift from customer-service roles toward sales roles. Training on AI tools also matters, but training alone is not enough. If the only message is “learn the tool or be replaced by it,” workers will hear a threat, not an invitation.

The deeper challenge is that firms and workers do not have identical interests. Companies will use AI to cut costs where they can. Employees know this. Governments will have to handle much of the social burden if AI causes broad disruption. Still, managers have agency over how much distrust they create inside their own organizations. They can make the transition feel arbitrary and extractive, or they can make it feel like a difficult but legible shift in which people get a fair chance to adapt.

The article’s takeaway is not that executives should hide the consequences of AI behind softer language. It is the opposite: they should speak plainly, but with precision and responsibility. Calling people “lower-value human capital” is not candor. It is a failure to understand what employees are afraid of. Better AI leadership starts with a more adult conversation: some tasks will be automated, some roles will change, some people may need new paths, and the employer has a duty to make those paths real before asking workers to trust the transformation.