9 reasons AI isn’t about to take your job (yet)

(SeaPRwire) –   Employers are facing immense pressure to adopt AI and reduce their workforce. Investors and CEOs envision significant cost reductions and increased profit margins; every Chief Information Officer is compelled to devise an AI strategy to keep pace with competitors. Visions of AI-agent-driven transformations are pervasive.

However, leaders should not feel obligated to hastily embrace a future that has not yet arrived. There are numerous reasons for exercising caution. Here are nine:

“Experts” have frequently been wildly inaccurate in their predictions. The Nobel laureate and AI pioneer Geoffrey Hinton stated in 2016, “People should cease training radiologists now… It’s simply undeniable that within five years, deep learning will outperform radiologists.” Yet, a decade later, few if any radiologists have been replaced. Google co-founder Sergey Brin promised in 2012 that driverless cars would be ubiquitous by 2017. Today, 14 years after that pledge (and many subsequent ones by Elon Musk), fully autonomous vehicles remain a limited experiment, available in only a small number of cities with favorable weather conditions.

Big Tech desires you to believe it has developed artificial general intelligence. That does not make it true. When technology CEOs warn of an employment Armageddon, they might be covering their bases in case such an event actually occurs, but then again, perhaps they merely wish for you to inflate their companies’ valuations. Approach every projection they make with a degree of skepticism.

When it comes to employment impact, the figures from AI giants do not support their assertions. Anthropic’s CEO has been cautioning about a job “apocalypse,” but Anthropic’s own recent research demonstrated the disparity between perception and reality. The company projects substantial potential for what AI might achieve in sectors like finance and architecture. However, what it termed “observed AI coverage” (a polite phrase for what is actually happening in the real world) constituted a comically small fraction of that theoretical scope. What they imagine AI might accomplish and what it is genuinely doing are vastly different.

Current AI is “jagged” (proficient in some areas but not others), meaning it can seldom entirely replace a human. AI can certainly enhance the productivity of some workers, but even on tasks that AIs are adept at, models and agents often make trivial errors, some of which are difficult to detect. And tasks are not equivalent to jobs: Even if AI can perform a segment of an individual’s role, it does not signify it can execute the entirety of that person’s job.

Current AI models still encounter difficulty extending beyond language. Some white-collar occupations involve only words, but many necessitate visual comprehension: interpreting images, charts, diagrams, blueprints, maps, and so on. It might seem straightforward to imagine AI assuming every role, especially if one perceives it as a form of magic. But once you recognize that current AI is a tool, with specific strengths and weaknesses, you begin to understand that this technology is only likely to displace workers in certain professions and not others (and more frequently will simply augment human jobs). Even in domains like customer service that might appear uncomplicated, results are often disappointing. The Remote Labor Index focused on jobs that could be accomplished entirely over the internet, and found that less than 4.5% could actually be adequately completed by AI agents.

Most physical labor extends well beyond what current AI can achieve. Do not anticipate AI replacing plumbers, carpenters, auto mechanics, nurses, house cleaners, forest rangers, chefs, appliance repair workers, gardeners, or many other jobs anytime soon.

Many layoffs that have been attributed to AI are not genuinely AI-related. This may have been the case for the recent extensive layoffs at fintech firm Block; some interpreted it as an effort by CEO Jack Dorsey to restore investor confidence after its stock declined. In many instances, AI may be serving as a pretext to conceal redundancies that are actually driven by financial underperformance or earlier overhiring.

Some layoffs that are attributed to AI do not last. I refer to this as the Klarna Effect, named after the buy-now, pay-later company Klarna, which proudly announced massive AI-driven layoffs only to reverse them. Many of the individuals dismissed worked in customer service, but after 11 months, Klarna concluded that (at least in specific cases) “real humans” were required after all.

The overall impact on productivity and return on AI investment has, to date, been modest. Every company is allocating resources to AI, but so far, the majority are not realizing substantial returns.

All this could change; it probably will someday—but most likely not until we witness more radical advancements in AI, which could be a decade or more away. In the interim, the advice is straightforward: Do not concentrate on substituting humans. Focus on how you can leverage AI to assist your existing workforce.

Gary Marcus holds the title of emeritus professor of psychology and neural science at NYU, and is the author of six publications, among them Taming Silicon Valley.

This piece is featured in the April/May 2026 edition of under the title “9 reasons not to panic (yet) about AI.”