Short answer: Both. In the next 5–10 years AI will cause large job churn—eliminating some roles while creating others. Whether it’s net-positive or net-negative depends on adoption choices, productivity gains, and how well workers and firms adapt. With strong reskilling and policies that spread productivity gains, the net can be positive; without them, expect more displacement and inequality.
What major studies say
•OECD (2019): ~14% of jobs at high automation risk; another ~32% will change significantly.
•Goldman Sachs (2023): About two-thirds of jobs have tasks that could be automated; up to 300M FTEs exposed globally, but potential for sizable productivity growth that could create new demand and jobs.
•WEF Future of Jobs (2023): By 2027, 69M jobs created and 83M displaced globally (net −14M), with big sector shifts.
•IMF (2024): AI will affect ~40% of jobs worldwide (advanced economies ~60%). Outcomes split between complementing and substituting labor; inequality risks rise if gains aren’t shared.
Where jobs are likely to grow
•Healthcare and care economy; skilled trades; green energy and infrastructure; data/AI, cybersecurity, and advanced engineering; product management; sales and customer success; education and training.
•Entrepreneurial opportunities building AI-enabled products and services.
Where jobs are most at risk
•Routine, rules-based, and document-heavy work: basic bookkeeping, data entry, some back-office operations, customer support scripts, simple coding, transcription, paralegal/contract review, and some marketing production tasks.
Key determinants of the net effect
•Whether firms use AI mainly to cut headcount vs to augment workers and launch new products.
•Speed of adoption and diffusion of productivity gains into lower prices, higher demand, and complementary hiring.
•Quality and scale of reskilling, labor mobility, and safety nets.
•Regulation and standards that encourage “task-creating” innovation, not just automation.
What individuals can do now
•Become AI-augmented: learn to use leading AI tools in your field.
•Build T-shaped skills: strong domain expertise plus data/AI literacy.
•Double down on human advantages: problem framing, judgment, interpersonal communication, creativity, leadership, and ethics.
•Move toward roles that design, supervise, and integrate AI systems (workflow design, QA, governance).
•Keep a portfolio mindset: continuous learning and willingness to transition.
Bottom line: Expect more displacement in the short run and significant reshuffling across occupations. Over the longer run, AI can create more and better jobs if businesses, workers, and policymakers steer it toward complementing people and expanding new kinds of work. Without that, the gains will be concentrated and net job losses or wage polarization become more likely.