This is a genuinely contested question among economists, and the honest answer is: it depends heavily on timeframe, measurement method, and which jobs/sectors you're looking at—and reasonable experts disagree. I don't think I can give you a confident single verdict, but let me lay out the main considerations.
The case for "takes away more"
•Direct displacement in specific tasks: AI is already automating routine cognitive work—first-pass drafting, basic coding, customer service scripts, data entry, routine analysis. Jobs or job tasks built around these functions are shrinking.
•Speed of this transition vs. past ones: Previous automation wayes (industrial revolution, computers) took decades to diffuse. Generative AI is being adopted across white-collar work much faster, which could mean less time for labor markets to adjust, retrain, and create replacement roles.
•Hits white-collar/knowledge work: Unlike earlier automation (mostly physical/manual labor), this wave affects higher-education, higher-wage jobs—lawyers, analysts, programmers, writers—which is a new dynamic with less historical precedent to reassure us.
The case for "creates more" (or at least net neutral)
•Historical pattern: Every major technology (steam power, electricity, computers, internet) initially destroyed specific jobs but created more new ones than it eliminated, plus entirely new industries nobody could have predicted in advance.
•Productivity → demand: When AI makes certain work cheaper/faster, it often increases overall demand for that type of output (e.g., cheaper legal research could mean more legal services consumed overall, not fewer lawyers).
•New job categories: Prompt engineering, AI oversight/auditing, AI ethics, data curation, human-AI collaboration roles—these didn't exist five years ago and are growing.
•Complementarity, not just substitution: Many jobs will be transformed rather than eliminated—AI augments human judgment, creativity, and interpersonal skills rather than replacing the whole job.
Why there's no clean answer
•It's not one labor market: Radiologists, truck drivers, customer service reps, software engineers, and artists face wildly different AI exposure and risk profiles.
•Net vs. gross effects: Even if AI creates as many jobs as it destroys in aggregate, the people who lose jobs are rarely the same people who get the new ones—causing real pain even in a "net positive" scenario.
•Depends on policy choices: How governments handle retraining, social safety nets, and regulation will significantly shape whether the transition nets out as broadly beneficial or creates serious inequality/displacement.
•Forecasts vary wildly: Serious economic institutions and researchers have published estimates ranging from mass job creation to significant net job loss, often depending on their assumptions about adoption speed.
My honest take: I lean toward thinking this will likely follow the historical pattern of net job creation over the long run (10-20+ years), but with a rockier, faster, and more unevenly distributed transition than previous technological shifts—meaning real short-to-medium-term pain for specific occupations even if aggregate numbers eventually look fine.
Would you like me to look into current data on specific sectors, recent studies, or economist forecasts to make this more concrete?