Will AI destroy more jobs than it creates?
Who says what
Researchers’ documented positions on “Will AI destroy more jobs than it creates?”
Documented positions only, coded with cited evidence by an AI agent: 26 of 400 mapped researchers so far. Everyone else is unknown: no reviewed evidence, which is neither agreement nor neutrality.
Agree (6)
Joseph Stiglitz agrees
Has stated that AI will replace people in more routine jobs and that, in his judgment, more jobs will be lost than created; his research describes AI as largely labor-saving.
Evidence
- “There will be jobs created, but my judgment is that there will be more jobs lost.” — source
- Artificial Intelligence, Globalization, and Strategies for Economic Development
Anton Korinek leans agree
His analysis of AI's implications for income distribution takes significant displacement and wage pressure seriously.
Christian Catalini leans agree
Argues that AI absorbs labor whose output can be measured, with wages for measurable tasks compressing toward the cost of compute and entry-level apprenticeship eroding; the paper does not give a net job-creation estimate.
Evidence
- Some Simple Economics of AGI
- “tasks that are highly complex but purely measurable will face rapid wage compression toward the marginal cost of compute” — source
Daniel Susskind leans agree
Argues that structural technological unemployment, where AI's substituting effect gradually outweighs its complementing effect, is possible though not inevitable, and that the threat of declining labor demand should be taken more seriously than before.
Evidence
- “Is that scenario inevitable? No, of course not. But is it possible – I argue that it is.” — source
- Technological Unemployment
Klaus Prettner leans agree
Models technological unemployment from automation in search-and-matching frameworks and finds displacement pressure real.
Sarah Bana leans agree
Documents deterioration in AI-exposed jobs, beginning even before ChatGPT.
It's complicated (11)
Anna Salomons it's complicated
Finds automation has displaced labor from the income share it once earned, even as employment kept growing — and that 'new work' offsets matter.
Bryan Seegmiller it's complicated
Co-authored work finds that tasks more exposed to AI see reduced labor demand, offset in part by workers reallocating effort to other tasks; related historical work finds new job categories have counterbalanced task-displacing automation.
Daron Acemoglu it's complicated
His task framework shows automation displaces labor while new tasks reinstate it — the net outcome depends on which kind of AI we build, which is why he pushes 'pro-worker AI'.
Georgios Petropoulos it's complicated
He frames automation as having opposing displacement and productivity effects, writing that displacement may dominate in the short run while job creation can dominate in the longer run; his EU robot study measures both effects.
Evidence
- “So, past cases suggest that in the short run the displacement effect may dominate.” — source
- The impact of industrial robots on EU employment and wages: A local labour market approach
Jakub Growiec it's complicated
His models find partial automation of complex tasks raises demand for human work in the remaining sub-tasks, while full automation lowers it; he links a shift of cognitive workers into manual jobs to AI becoming capable across cognitive tasks.
Joshua Gans it's complicated
Titles his analysis 'the ambiguous labor market impact of automating prediction' — displacement and augmentation both plausible.
Maria del Rio-Chanona it's complicated
Freelance-platform evidence finds falling demand for AI-substitutable skills but rising demand in complementary and unaffected clusters.
Neil Thompson it's complicated
His research models automation as replacing some tasks while raising or lowering the expertise needed in remaining ones, and his cost-effectiveness study expects displacement to be substantial but gradual, with many jobs reshaped rather than eliminated.
Pamela Mishkin it's complicated
Co-authored study finds freelance demand for skills substitutable by generative AI, such as writing and translation, fell 20-50% relative to trend, while demand rose in complementary or unaffected skill clusters.
Pascual Restrepo it's complicated
Automation displaces labor while new tasks reinstate it; he warns the current path of AI leans toward displacement ('the wrong kind of AI').
Simon Johnson it's complicated
Co-authored papers state AI may replace many workers and that its automation capacity is substantial, while also arguing its potential to create new tasks and extend human judgment is equally large; no net job forecast is given.
Disagree (9)
Christos Makridis disagrees
Finds employment rises about 4% where AI complements human work and shows no significant effect where AI can perform tasks independently, and writes that AI is not destroying jobs.
Evidence
- AI, Output, and Employment
- “AI is not destroying jobs.” — source
Prasanna Tambe disagrees
A co-authored op-ed argues AI is more likely to add tasks than eliminate jobs and predicts it will create more jobs than it eliminates, as past technologies have.
Evidence
- “But if they do, it will create more jobs than it eliminates, as has virtually all technology in the past.” — source
Ajay Agrawal leans disagree
Argues for a 'Turing Transformation' in which AI augments rather than replaces workers, raising demand for human skills.
Alexander Oettl leans disagree
Co-authored task-based models in which AI enhances worker productivity without automating tasks, citing evidence that contemporary AI is more often used to augment workers than to replace them.
Christopher A. Pissarides leans disagree
He has argued that work is not finite and that labor markets will adjust to AI over time, with upskilling as the main need, rather than predicting net job loss.
Evidence
- “There's just no limit to how much work humanity could generate if they really want to work” — source
David Autor leans disagree
His research shows most of today's jobs are 'new work' created since 1940 — technology has historically created more work than it destroyed, and current AI effects on aggregate employment remain undetectable.
Diane Coyle leans disagree
Expects some labor-market disruption but not the large-scale job loss described by some AI industry leaders.
Evidence
- “I don't myself think it will be a job apocalypse of the kind that some of the AI leaders have been talking about.” — source
Erik Brynjolfsson leans disagree
Emphasizes augmentation over automation (the 'Turing Trap' is a choice, not destiny): AI restructures work and creates new tasks rather than simply destroying more jobs than it creates.
Lynn Wu leans disagree
Presents research finding that firms investing in robotics often hire more people overall, reducing managers rather than frontline workers, with middle-skill roles declining and high- and low-skill roles growing. The evidence concerns robotics adoption rather than AI broadly.