AI job paradox: Occupations seeing greatest losses also see greatest gains - here's why

akinbostanci/ E+ via Getty Images

Follow ZDNET: Add america arsenic a preferred source on Google.

ZDNET's cardinal takeaways

• Key IT roles are being scaled backmost and ramped up for the AI era. 
• IT operations, bundle development, and cybersecurity are seeing some cuts and gains.
• A 'reorganization of work' is reshaping occupation roles. 

A caller manufacture survey shows a scope of cuts crossed respective IT occupation categories -- but besides a boost successful hiring for the aforesaid types of positions.  

The survey of 2,050 executives crossed the globe, released by Snowflake, reveals occupation losses owed to AI, portion besides showing that the aforesaid occupations are benefiting from the AI trend. For example, 40% of executives surveyed study cuts to IT operations owed to automation, but 56% study further hiring for these positions. Another 26% are seeing cuts successful bundle improvement jobs, but 37% besides study accrued hiring. For information analysts, the divided is 37% and 37%.

Also: Job hunting? 5 ways you tin basal retired successful 2026 - and bushed AI screening tools

For jobs extracurricular of IT, the representation is much straightforward -- and for most, little dramatic, with the objection of lawsuit work and enactment staff. The lawsuit work workforce declined by 37% among the organizations surveyed; portion lone 15% were expanding hiring, (You could blasted AI, but outsourcing whitethorn beryllium different culprit down the cuts.) For manufacturing and supply-chain operations, 6% were cutting, portion 13% were hiring. For selling staff, 16% were cutting versus 12% hiring. 

The comparisons among these rising and declining occupation roles are not precisely apples-to-apples. "What we're seeing is simply a reorganization of work, not a elemental enlargement oregon contraction of headcount," Baris Gultekin, vice president of AI astatine Snowflake, told ZDNET. "AI is taking implicit repetitive, manual tasks wrong these roles. At the aforesaid time, it's creating wholly caller responsibilities astir AI integration, governance, information engineering, security, and show oversight. In that sense, it's not arsenic black-and-white arsenic companies conscionable cutting oregon adding jobs. They are reshaping the jobs themselves to enactment caller AI workflows." 

Evolution not elimination

Asked whether generative AI had driven occupation creation, occupation loss, oregon some astatine their organizations, 42% replied lone that jobs person been created by gen AI, portion 11% indicated that jobs person been lost. Another 35% study that jobs person some been created and mislaid owed to AI. The remaining 13% said that AI had not affected their employment 1 mode oregon the other.

Overall, 77% reported immoderate occupation creation, with oregon without accompanying occupation loss, the probe finds. "That uncovering indicates that this is little astir elimination of roles, and much astir evolution," Gultekin said.    

Also: Need a caller job? These AI roles are the fastest increasing successful the US, says LinkedIn

 "As soon arsenic AI moves beyond experimentation, the accomplishment requirements shift," helium explained. "Running a aviator is 1 thing. Operating AI astatine standard wrong an endeavor is thing other entirely. It requires beardown information foundations, wide governance models, infrastructure expertise, and radical who recognize however to monitor, evaluate, and optimize exemplary show implicit time."

The survey shows that 35% of organizations mention accomplishment gaps arsenic a large obstruction to AI success. "That's a wide awesome that the constraint is nary longer conscionable AI technology, but the expertise needed to guarantee its occurrence successful the enterprise," Gultekin said. "As companies determination toward much precocious agentic usage cases, the request for oversight grows. Someone has to guarantee information quality, negociate risk, and marque definite these systems are acting responsibly. In that sense, AI does not region the request for people, but it does alteration expectations astir what those radical request to know."

Increased request successful higher-skill areas

The information suggests that the existent communicative astir AI usurping exertion jobs "is much analyzable than galore think," helium continued. "Historically, large exertion shifts alteration the creation of enactment much than they trim full employment. We are seeing a akin signifier with generative and agentic AI. Some task-based roles are being automated. At the aforesaid time, request is increasing successful higher-skill areas specified arsenic AI operations, cybersecurity, information engineering, and governance."

Also: The biggest AI threats travel from wrong - 12 ways to support your organization

In addition, greater acquisition with AI translates into occupation growth. "Organizations that are further on successful AI adoption are much apt to study a nett affirmative employment impact, Gultekin notd. "That is an important information point. It suggests that alternatively of an outright illness successful jobs, what we're really seeing is endowment reallocation toward much strategic, technical, and AI-enabled roles."

Among different topics, the Snowflake survey explored the starring concern and method concerns associated with agentic AI improvement and deployment. Leading concerns see interoperability issues (42%), bequest strategy incompatibility (39%), providing real-time information processing for cause determination making (42%), occupation displacement (29%), maintaining quality oversight/preventing rogue actions by agents (29%), and concerns implicit information retention and usage (29%).