Will AI create or take away more job opportunities?

[1][2]AI is widely expected to create more job opportunities than it eliminates on net, but it will displace or transform many existing roles, requiring large-scale worker mobility, reskilling, and adaptation. The outcome is not predetermined—it depends heavily on the pace of adoption, how businesses redesign work (augmentation vs. pure replacement), education/training systems, and public policy.[1]

Major Forecasts Point to Net Job Creation

Other analyses (e.g., Goldman Sachs) show mixed short-term effects: substitution losses partially offset by augmentation-driven gains and firm expansion, with modest overall labor market drag so far and bigger impacts on younger/less-experienced workers.[5]

Current Evidence: Transformation More Than Mass Destruction (So Far)

As of the mid-2020s, there is limited evidence of economy-wide AI-driven mass unemployment. Unemployment in highly AI-exposed jobs has not diverged sharply since generative AI tools like ChatGPT emerged. Effects appear more in:

IMF estimates suggest ~40% of global jobs (and ~60% in advanced economies) are exposed to AI in ways that could augment, downgrade, or eliminate them—higher in cognitive/professional work in rich countries, lower in manual/agricultural work in developing ones. Exposure does not equal automatic loss; complementarity (AI boosting human productivity) is common.[9]

How AI Creates and Takes Jobs

Displacement/risk areas: Routine cognitive and administrative tasks (data entry, basic analysis, certain customer service, some content generation, bookkeeping). Generative AI particularly affects knowledge work that was previously hard to automate.

Creation and growth areas:

Most jobs will likely be transformed rather than eliminated outright—workers shift time toward higher-value tasks (judgment, creativity, interpersonal skills, complex problem-solving) while AI handles routine ones. Google’s chief economist and others emphasize redesigning workflows for the biggest gains.[10]

Historical Context

Past general-purpose technologies (mechanization of agriculture, electricity, computers, the internet) caused major short- and medium-term displacement but were net job creators over longer periods, raising productivity and living standards. Agriculture employment plummeted as a share of the workforce, yet overall employment and new industries grew. Many of today’s jobs did not exist decades ago. AI follows a similar pattern in principle but may act faster on cognitive tasks and create different distributional challenges (potentially affecting higher-education white-collar work more than prior waves).[11][12]

Recent research notes that technology has sometimes automated more jobs than it created in certain periods (e.g., post-1980 in some analyses), with new work often favoring younger, more educated workers. AI’s ultimate balance remains uncertain and depends on implementation.[13]

Key Uncertainties and Risks

Bottom Line and Path Forward

AI is more likely to create a net surplus of job opportunities while profoundly changing the nature of work, consistent with major institutional forecasts and historical patterns. The bigger near-term story is disruption, occupational switching, and the need for mobility rather than absolute job scarcity.[2]

Success hinges on deliberate action: massive investment in lifelong learning and AI literacy + human skills (critical thinking, creativity, collaboration, adaptability); businesses redesigning processes around human-AI teams rather than just cutting headcount; governments supporting transitions (training, portable benefits, education reform); and inclusive policies so gains are broadly shared. Individuals who build AI fluency and complementary skills will be best positioned. The technology itself does not dictate the outcome—human choices about deployment and support systems do.