This is one of the most debated economic questions of our time, and the honest answer is: it's genuinely uncertain, with legitimate evidence and expert opinion on both sides. Let me lay out the key considerations rather than giving a false sense of certainty.
The case for "more jobs destroyed"
•Direct automation risk: Roles involving routine cognitive work (data entry, basic analysis, first-draft writing, customer service, paralegal work, coding) are increasingly automatable by current AI systems.
•Speed of change: Unlike past technological shifts (decades), AI capability is improving rapidly, potentially outpacing workers' ability to retrain.
•Winner-take-most dynamics: A small number of highly productive AI-augmented workers or firms could do the work that previously required many people.
The case for "more jobs created"
•Historical pattern: Every major technology (steam power, electricity, computers, internet) initially displaced specific jobs but ultimately created more jobs than it destroyed — often in roles that didn't previously exist (web designer, app developer, data scientist).
•New categories of work: AI is already spawning new roles — prompt engineers, AI trainers/evaluators, AI ethics specialists, human-AI collaboration managers.
•Productivity → demand: When AI makes things cheaper/faster to produce, demand for those goods/services often rises, requiring more total labor (including in adjacent roles) — this is the classic "Jevons paradox" dynamic.
•Complementary tasks: Many jobs have components AI can't easily do — judgment calls, physical dexterity, emotional labor, trust-building, novel problem-solving — so AI may augment rather than replace many workers.
Why this time might be different (or might not)
•AI is a general-purpose technology affecting cognitive labor broadly, not just one sector — some economists argue this makes comparisons to past automation waves (which were narrower) less reliable.
•Counter-argument: even general-purpose technologies like electricity touched nearly every industry and still led to net job growth over time, just with painful transitions.
The more likely real outcome
Most economists expect net job numbers to stay roughly stable or grow over the long run, but with:
•Significant short-to-medium-term disruption — specific jobs/industries losing workers faster than new ones absorb them
•Uneven effects by skill level, geography, and industry — some sectors hit hard, others barely touched, some booming
•A shift in what jobs look like rather than a simple net increase or decrease — more emphasis on skills AI can't replicate
Bottom line: I don't think anyone — including top economists, AI researchers, or tech CEOs — can confidently predict the exact net effect. The pattern from history leans toward "more jobs created over time," but the pace, fairness, and distribution of that transition is the real open question, and it's one where policy choices (retraining programs, safety nets, education) will matter a lot in determining who wins and who loses.
Would you like me to dig into current data or expert forecasts on this (e.g., recent reports from the WEF, McKinsey, or labor economists)?