AI's impact on jobs is the big question — I'll weigh the latest forecasts on what it creates versus what it replaces.
On balance, most major institutional forecasts expect AI to both create and destroy jobs - but create slightly more in the net over the next 5-10 years. The gains and losses do not fall on the same people, places or skills, which is why the transition is expected to be painful even if the total number of jobs rises.
1. What the headline forecasts say: net gain
•World Economic Forum – Future of Jobs Report 2025: Projects 170 million jobs created and 92 million displaced by 2030, a net gain of 78 million. The same report is summarized elsewhere as 92 million roles displaced and 170 million created globally by 2030. The report notes that figure covers all forces reshaping work, not AI alone.
•McKinsey Global Institute (U.S. focus): AI and automation will cut demand for about 36 million U.S. jobs by 2035 while growth elsewhere creates about 41 million. The same forecast is described as reducing demand for roughly 36 million jobs while simultaneously generating about 41 million new roles. Net result is about 5 million jobs ahead.
•Gartner HR Research (May 2026): Beginning in 2028, AI will create more jobs than it eliminates, and in a longer formulation: Beginning in 2028, AI is expected to generate more jobs than it eliminates. Gartner adds that until then the impact is neutral through 2026, with more evident gains in 2028-2029.
In that sense the institutional consensus is: short-term losses, medium-to-long-term net creation.
2. Why both happens at once
AI does not replace whole occupations at once, it replaces tasks – reading documents, extracting information, filling forms, drafting routine correspondence, handling first-contact tickets.
•Goldman Sachs Research (2023) said workflow changes could expose the equivalent of 300 million full-time jobs globally to automation. A 2026 Goldman Sachs measure found about 16,000 net U.S. job losses per month from AI, with employment for software developers aged 22 to 25 falling nearly 20% since 2024.
•IMF (2024) put almost 40% of global employment in the exposed group, and about 60% in advanced economies. Other IMF summaries put it as almost 40% of global employment is exposed to AI, rising to about 60% in advanced economies.
Routine clerical, administrative, customer service, junior coding, bookkeeping and basic translation tasks go first, while new demand appears in:
•AI itself – building, training, deploying and governing AI systems
•Healthcare, construction, management and data centres – McKinsey notes the shrinking work is in office and administrative support, retail and transportation, much of it lower-paid, and the growth is in healthcare, construction and management
•Augmented roles where humans direct AI – judgment, trust and decision-making work that Gartner says future jobs will focus on
Historical parallel often cited: the ATM was supposed to kill bank tellers — instead cheaper branches meant more branches and tellers became advisors.
3. The catch: distribution is unequal
Forecasts warn a net gain says nothing about whether a displaced worker can move into a new role without training.
•McKinsey's base case: roughly 11 million workers (about 7% of the workforce) would need to leave their occupation entirely. About 770,000 workers a year would need to switch into an entirely different field... about 3.6 times the historical average.
•Only one in seven displaced workers has a direct path into a growing job that needs little retraining and pays at least as much. Nearly half face what the report calls an "unpaved" path, blocked by large skill gaps or credential requirements, and about 85% of growing jobs require a credential.
•The chance of having to change occupations is 7.6 times higher for lower-wage workers than for higher-wage ones.
•Public expectation is pessimistic: In 34 of 37 countries surveyed by Pew in September 2025, more people expect AI to eliminate jobs than create them; in the U.S., 71 percent of adults told Pew that AI will lead to fewer jobs over the next twenty years.
Early evidence of churn: 86% of administrative and clerical workers at highest risk are women and 79% of employed U.S. women work in jobs at high automation risk versus 58% of men in Brookings data cited in analysis, and 37% of business leaders say they plan to replace human workers with AI by the end of 2026.
4. Short term vs long term
•2026-2028: Net displacement in entry-level and routine white-collar work. Goldman measured about 16,000 net losses per month in the U.S. in early 2026. Hiring for AI-exposed junior roles is down ~15% year-over-year.
•2028-2032: WEF's 92 million displaced figure begins to materialize, but reskilling demand spikes – over 40% of workers worldwide will require significant upskilling by 2030 per IMF framing, and 40% of core skills are expected to change by 2030 per WEF survey.
•2032+: Two paths: a Jevons Paradox-style expansion where cheaper output creates more demand (cheaper legal work = more people afford lawyers), or concentration of gains among AI owners leading to precarity and need for policies like UBI.
> Gartner's summary captures the consensus view: “AI is ultimately going to result in more job gains than losses, but in the process it's going to break down millions of careers”.
Bottom line
Will AI create or take away more jobs?
•In raw numbers, most models say create more than it destroys by around 2028-2030 – WEF +78 million globally by 2030, McKinsey +5 million in the U.S. by 2035.
•In human terms, it will take away more existing jobs than it preserves – 36 million U.S. jobs worth of demand cut, 92 million roles displaced globally – and create new jobs that require different skills, credentials, and often locations.
The challenge is not scarcity – “The next decade's challenge is mobility, not scarcity” – but whether workers, especially lower-paid clerical, retail and support workers, can cross into those new roles.