Short answer: both. AI will eliminate some tasks (and a subset of jobs built mostly on those tasks), while creating new tasks, roles, and even industries. Whether there are “more” jobs in total depends on timing, policy, and how quickly organizations reinvest productivity gains into growth.
What tends to shrink
•Roles heavy on routine, predictable knowledge work: basic data entry, transcription, simple customer support, straightforward bookkeeping/reporting, templated copy, some QA/testing, and portions of junior programming or design work.
•Middle layers of work that mainly transform inputs to outputs without much context-setting, relationship-building, or physical presence.
What tends to grow
•AI-related technical roles: ML engineering, data platform/MLOps, AI product management, prompt/tooling development, evaluation, safety, security, and governance.
•Complementary “human-in-the-loop” roles in healthcare, education, sales, customer success, law, finance, and operations—where judgment, trust, and accountability matter.
•Compliance, risk, and audit for algorithmic systems.
•Skilled trades and in-person services that are hard to automate end-to-end, often augmented by AI planning or diagnostics.
Timeframe matters
•Near term (next 3–5 years): noticeable task automation and some net job losses in specific functions; many jobs get redesigned rather than removed. Big premium on learning to use AI to boost your own throughput.
•Medium term (5–10 years): if firms channel productivity gains into new products and markets, demand can expand and offset earlier losses, often with different skill mixes.
•Long term: historically, general-purpose technologies raise total employment, but the transition can be uneven across regions and education levels.
What ultimately decides net outcomes
•Adoption speed and capability: how good, cheap, and reliable AI becomes for specific workflows.
•Elasticity of demand: whether lower costs translate into more volume or just fewer workers.
•Reskilling capacity: how quickly workers can move into complementary roles.
•Regulation and trust: rules that shape where AI can be deployed and how.
•Organizational design: whether leaders use AI to cut headcount or to grow.
How to stay on the opportunity side
•Become the person who can frame problems, not just complete tasks; learn to decompose work and orchestrate AI tools.
•Build domain expertise plus data/AI literacy; document AI-augmented workflows in a portfolio.
•Double down on interpersonal skills: negotiation, client management, storytelling, ethics, and accountability.
•Get comfortable with basic automation (APIs, low-code, scripting) to connect tools and eliminate drudge work.
•Track emerging standards in your industry (privacy, model governance, accuracy thresholds).
Bottom line: Expect AI to reallocate work more than it simply “removes jobs.” In the short run, some roles shrink; in the medium run, new roles and higher-output teams can more than offset losses—if individuals and organizations invest in adaptation. If you tell me your field and experience level, I can give a tailored map of tasks to automate, skills to learn next, and roles likely to grow.