The Public Does Not Believe the AI Jobs Promise
A global survey shows that most people, in most countries, expect AI to shrink employment, and leaders who ignore that will meet resistance.
Sign up for my Substack daily AI newsletter here.
See my AI Training course portfolio for corporate Business Leaders here.
Follow me on LinkedIn: https://www.linkedin.com/in/johanosteyn/
For years, the companies building artificial intelligence have promised that it will free us from drudgery, create new kinds of work, and lift everyone’s productivity. The public is not buying it. A major global survey released recently found that in country after country, most people expect AI to destroy more jobs than it creates. This is not a fringe worry or a passing mood. It is a settled, near-universal verdict, and it now shapes how workers, customers and regulators respond to every AI project a company launches. For business leaders in South Africa and elsewhere, the temptation is to treat this as a communication problem to be managed with better messaging. That would be a serious mistake. When people have already decided what AI means for their livelihoods, the task is not to persuade them harder, but to earn their trust.
CONTEXT AND BACKGROUND
The scale of the scepticism is striking. In thirty-four of the thirty-seven countries surveyed by the Pew Research Center, people were more likely to say AI will lead to fewer jobs than more, and in wealthy nations such as the United States, Australia and South Korea, around seven in ten adults expect job losses over the next twenty years. In the United States that share has risen sharply in only two years, and pessimism climbs with national income, so the economies best placed to profit are often the most anxious.
Trust in the firms deploying AI is falling in step. In the same country, Gallup found that nearly eight in ten adults now expect AI to reduce the number of jobs over the next decade, and only about a quarter say they trust businesses to use it responsibly, a figure that has reversed two years of gradual improvement.
This public verdict sits awkwardly against the views of those who build the technology. The Stanford AI Index records a persistent gap between experts, who remain markedly more optimistic about jobs, and a public that is not.
INSIGHT AND ANALYSIS
It would be comforting to dismiss this as public ignorance. That reading does not survive contact with the evidence, because the people running companies broadly agree with the pessimists. In the World Economic Forum’s own survey, more than half of executives expected AI to displace existing jobs and only about a quarter expected it to create them, while far more anticipated fatter profit margins than higher wages. The public is not being cynical. It is taking executives at their word and believing the financial reports over the town halls.
That is why the real obstacle to AI is no longer technical, but a matter of trust. A workforce that believes automation is coming for it will resist, disengage or quietly sabotage the very projects meant to raise productivity. In practice, this looks like shadow compliance, knowledge-hoarding and expensive enterprise licences sitting idle while morale erodes. Customers grow wary of firms they see automating people out of a living. Regulators, reading the same polls, feel pressure to act. Sentiment has become a business variable, and a leader who treats it as noise is managing only half the problem.
I have previously written about this, arguing that the question of which South African jobs should never be handed to a machine is one we must answer deliberately rather than leave to the market.
IMPLICATIONS
Nowhere are the stakes higher than in South Africa, where youth unemployment already sits near forty-six per cent. In wealthier economies, displaced workers can often absorb into gig work or move sectors. Here, with no meaningful safety net and social infrastructure already stretched, automation does not mean friction before a career transition; it risks permanent disenfranchisement. That turns board accountability from good ethics into sovereign necessity.
The response cannot be a better slogan. It has to be honesty about what will change, and a visible commitment to the people affected: serious investment in reskilling, a share of the productivity gains reaching wages and not only margins, and workers brought into decisions about automation rather than told about them afterwards. This is cognitive governance applied to the workforce, the discipline of putting human agency and labour transition on the board’s own audit list rather than delegating them to HR after the fact.
CLOSING TAKEAWAY
The most important finding in these surveys is not a number but a verdict. The world has looked at the promise that artificial intelligence will be good for work, and it does not believe it. That disbelief is a warning and an opportunity at once. Leaders who hear it as a public relations nuisance will spend years fighting their own workforces and customers. Those who treat it as an honest signal, and respond by making AI work for people and not only for profit, will earn something no campaign can buy. The promise was always ours to keep, not merely to make.
Author Bio: Johan Steyn is a prominent AI thought leader, speaker, and author with a deep understanding of artificial intelligence’s impact on business and society. He is passionate about ethical AI development and its role in shaping a better future. Find out more about Johan’s work at https://www.aiforbusiness.net
Comments