This is a genuinely contested question among economists, and the honest answer is: it depends on the time horizon, the sector, and the type of AI advancement we're discussing. There isn't a clean consensus, but I can lay out what the evidence and major perspectives suggest.
The case for net job loss (at least in certain segments)
•Routine and entry-level cognitive work is most exposed. Jobs involving repetitive data processing, basic coding, customer service scripts, and routine writing/analysis are already being automated or augmented heavily.
•Historical automation parallels (manufacturing automation, ATMs, self-checkout) show that specific tasks get eliminated quickly, even when total employment in an industry doesn't collapse.
•Some recent studies point to early labor-market effects concentrated in junior/entry-level roles, since AI can replace the kind of work new graduates traditionally did to build experience, while more senior roles that require judgment, relationships, and accountability are harder to automate.
The case for net job creation (or at least transformation)
•Every major technological shift historically created more jobs than it destroyed, though often not in the same roles or for the same people. Electricity, computers, and the internet all eliminated huge categories of jobs while creating new industries (e.g., "social media manager" didn't exist in 2000).
•New job categories are already emerging: prompt engineering, AI auditing/safety, AI trainers/data curators, human-AI workflow designers, and oversight/compliance roles.
•Productivity gains can expand overall economic activity, increasing demand for goods and services, which can create jobs elsewhere even in sectors not directly touched by AI.
•Many economists argue AI will mostly change job composition rather than cause mass unemployment — automating specific tasks within jobs rather than eliminating entire occupations wholesale.
Why this is hard to predict confidently
1.Speed matters. If AI capabilities advance faster than the economy can create/adapt new roles, there could be painful transition periods even if long-run job totals are fine.
2.Policy response matters enormously — retraining programs, education systems, social safety nets, and how fast regulation adapts will shape outcomes as much as the technology itself.
3."Jobs" vs. "tasks" distinction — most jobs involve bundles of tasks, and AI typically automates some tasks within a job rather than the entire job, which complicates simple "X jobs lost" predictions.
4.Generative AI is different from past automation because it affects cognitive and creative work, not just physical/routine labor — we don't have a perfect historical analogue.
My honest take
I don't think there's a universally correct answer right now — credible experts genuinely disagree, partly because it depends on policy choices we haven't made yet and capabilities we haven't seen yet. If you want, I can look into more recent data/reports (e.g., from WEF, McKinsey, or labor economists) on this, since there's been a lot of fresh research in 2024–2026 that might sharpen the picture. Want me to pull that up?