New data shows that AI isn't necessarily bulldozing jobs and entire departments, but making some jobs irreplaceable by renovating them one room at a time.
But I have written stories before on the jobs AI can’t replace.
Saying that “data jobs” are growing or disappearing might be essentially meaningless when some roles within the same department are soaring while others are contracting.
The safest career might not necessarily be the job AI cannot touch.
Hiring Is Down 24% The Visier data also arrives during a difficult period for job seekers.
New data shows that AI isn't necessarily bulldozing jobs and entire departments, but making some jobs irreplaceable by renovating them one room at a time. getty
Artificial intelligence has fueled predictions that millions of workers could eventually find themselves competing with machines for their jobs. But I have written stories before on the jobs AI can’t replace. Now, new workforce data suggests something more complicated is happening inside companies. Instead of simply replacing entire occupations, AI appears to be quietly reshuffling the value of different jobs, and making some distinctly human roles more important.
New research from workforce AI intelligence company Visier—based on more than 3.6 million employee records—finds that organizations are changing the workers they hire, which jobs they prioritize and where they deploy talent as artificial intelligence spreads through the workplace.
Among the most striking findings: Overall hiring rates have dropped 24%, while occupations including educators, lawyers, architects and security professionals emerge among the jobs that remain difficult to replace.
There’s another important shift taking place. Although younger workers still account for the largest proportion of new hires, companies are increasingly prioritizing experienced, mid-career employees between ages 35 and 50.
If the trajectory continues, that group could eventually represent the largest share of new hires. The findings suggest that the AI employment story may be less about entire industries disappearing and more about a massive reorganization taking place beneath the surface of companies.
AI Is Changing Jobs Inside Companies Visier analyzed workforce changes between 2022 and 2026 across more than 155 enterprise organizations and 30 job domains. One of its most important discoveries is that looking at industries alone can hide what AI is actually doing to employment. A bank and a hospital might operate in completely different industries, for example, but both employ people in finance, human resources, data analytics and other comparable functions. So Visier examined changes according to job domains and skills rather than simply comparing one industry with another. The differences can be dramatic. Data and Analytics increased its share of organizational headcount by 49% during the period studied, while Product Management increased 39%. Meanwhile, HR's share declined 3% and Finance dropped 4%. But even those numbers don’t tell the complete story. Looking deeper within Data and Analytics and two jobs traveling in opposite directions emerge. The share of AI engineers hired increased 251%, while the share of data scientists fell 32%, according to Visier. That’s an important distinction for workers. Saying that “data jobs” are growing or disappearing might be essentially meaningless when some roles within the same department are soaring while others are contracting. AI isn't necessarily bulldozing entire departments. It's renovating them one room at a time.
Why Some Jobs Are Becoming More ‘Irreplaceable’ Some occupations are resistant to AI replacement, and they offer another clue about how workers can protect their careers. Educators, lawyers, architects and security personnel involve responsibilities that cannot always be reduced to generating information quickly. They require judgment, accountability, interpersonal understanding, interpretation and, in many cases, human trust. AI can help lawyers research cases, but clients still need someone who understands nuance, strategy and the consequences of a legal decision. AI might help architects generate concepts, but architects must weigh safety, regulations, client needs and real-world conditions. Teachers can use AI to create lesson plans, but education involves motivating students, reading emotional cues, managing classrooms and understanding why one child isn’t learning the same way another child is. Security professionals must often evaluate ambiguous situations in real time, where context and human judgment matter. That distinction could become increasingly important for workers deciding where to invest their skills. The safest career might not necessarily be the job AI cannot touch. A more resilient career is one in which AI performs pieces of the job but cannot assume responsibility for the entire outcome.
Hiring Is Down 24% The Visier data also arrives during a difficult period for job seekers. Hiring rates have fallen approximately 24%, suggesting companies are becoming more selective about when and where they add employees. The decline doesn’t necessarily mean AI alone caused that contraction. Organizations have also faced economic uncertainty, restructuring and changing workforce strategies during the four-year period. Still, AI provides companies with another option that didn't exist at the same scale several years ago: Instead of automatically hiring another employee when workloads increase, employers can ask whether technology can absorb some of that work. That changes the equation. It can also raise expectations for the workers companies do hire. Instead of hiring five people to perform five narrowly defined functions, employers might increasingly hire fewer workers capable of using AI to accomplish more. And that could help explain another unexpected development in Visier’s data. The mid-career worker could be making a comeback, according to the study. Employees between 35 and 50 are increasing their share of new hires, Visier finds. Younger workers still account for the largest group today, but mid-career professionals could eventually overtake them if current patterns continue. Experience suddenly matters differently when AI can supply information. A machine can retrieve facts in seconds. What it doesn’t possess is 15 or 20 years of experience recognizing when those facts don’t fit the situation. That’s where veteran workers can have an advantage.
