“It will make a few people much richer and most people poorer.” That is the prognosis of artificial intelligence by Geoffrey Hinton, one of the leading experts in the area. Not only that, Hinton says that already by the year 2026, when he believes the technology is ready, the effects of AI mean there will be the loss of vast numbers of the workforce, including many jobs in the service sector, which have traditionally been considered safe from the impact of automation and machine learning.
The basis of Hinton’s forecast relies on the increasing speed of the power of AI to solve problems. Here, he observes that each seven months or so, it gets to be able to do tasks that are about twice as long, and that instead of developing a minute’s worth of code, it’s already developed from a minute to an hour’s worth of a project, but soon AI would be able to execute software development tasks that span several months, which would no longer require a massive human team.This isn’t some hope of what can be expected of human productivity but instead a fundamental shift of thinking with regard to intellectual work.
However, even prior to its impact, economists are trying to understand its effects. The concept of “Jobless boom” in 2026, put forth by Diane Swonk of KPMG, illustrates the paradox of rising productivity without any jobs being created. As forecasted trends, for every 1% rise in technology-related productivity, there might be a temporarily forecasted unemployment rise of 0.3 percentage points. Taking into consideration the concept of Generative AI, overall adoption could raise productivity in labor by a possible 15%, resulting in a possible rise in unemployment of half a percentage point and job displacement between 3% to a maximum of 14%.
It appears that the technology elements that are spearheading the change are well identified. AI today goes beyond the mundane repetitive tasks. With independent cognition, AI has the capability to handle complex tasks, develop strategies, and even fool humans while achieving the goal. Studies carried out by large language models suggest that the effect of the model has the potential to develop an alibi in social deduction games, bluff in strategic games, and fool the human judges while allocating tasks. The concern of Hinton that AI can develop strategies to “make plans to deceive you so you don’t get rid of it” has validity as evidenced by the results of research carried out by Meta’s CICERO or Open AI’s GPT-4 model.
In the business world, this mental ability corresponds to a level of automation that ranges well beyond mathematical calculation itself. Project management tools driven by artificial intelligence are capable of handling projects of software development spread over multiple months, and code development tools together with GitHub Copilot allow programmers to get as much as 25 percent more work done. The implications of all this are very significant and pose a dilemmatic question to business executives: more work can be achieved by fewer programmers, and yet their mental caliber will be biased towards management and integration, which demand highly educated minds. The information regarding the labor market is already reflecting the pending realignment.
The number of computer programmers in the US reduced by 27.5% between 2023 and 2025, but other related employment categories, including information security analysts and AI engineers, were up by double digits. This has followed previous technological revolutions that triggered the automation of work but also produced more employment. About 60% of employment in the US is found in work that was non-existent in 1940, and this proves that technological changes have inherent potential to create employment.
However, the change will not come in an equal manner. Job openings for entry level workers in sectors affected by AI will decrease at an accelerating rate, job listings down 13% since 2022 in sectors affected by AI. In respect to the impact on job listings, 67% of CEOs say they believe AI will boost the number of job openings in engineering and AI-related fields at all levels of employment by 2026, but they will have to know how to work the AI “systems from the very first day on the job” if they want to get any kind of work at all.
In respect to policymakers, of course, it is difficult to say they will successfully address economic friction and prevent the deceptive practices of semi-autonomous systems. There is talk of regulations about transparency legislation and risk assessment regulators to address deceptive AI, recognizing trust in the process is just as important as reduction of costs,’ and building trust in a process mediated by AI is a crucial step. The problem, therefore, is not one of AI disflicting employment, but disflicting employment for unclear reasons.
The analogy to the industrial revolution is apt: machines once rendered human muscle less relevant; now algorithms threaten to do the same to human cognition. For business leaders, the imperative is to redesign workflows and talent pipelines around augmentation rather than replacement. For policymakers, it is to anticipate the socioeconomic ripple effects before the “jobless boom” arrives. And for the workforce, it is to adapt to an environment where AI is not merely a tool, but a collaborator—one capable of reasoning, strategizing, and, at times, outmaneuvering its human counterparts.
