Anthropic CEO: “50% of all entry-level Lawyers, Consultants, and Finance Professionals will be completely wiped out within the next 1–5 years." grad students and junior hires are cooked. https://t.co/LGpi7kUl1y
This WEF infographic (Fastest growing and declining jobs by 2030) lists occupations expected to decline — including 'legal officials', 'legal secretaries', 'bank tellers', and 'accounting, bookkeeping and payroll clerks' — illustrating that legal and finance entry-level roles are among those employers expect to shrink, which is directly relevant to the Anthropic CEO’s warning about AI-driven losses for junior lawyers and finance professionals.
Source: World Economic Forum (Future of Jobs Report 2025)
Research Brief
What our analysis found
Anthropic CEO Dario Amodei warned in a May 28, 2025 Axios "Behind the Curtain" interview that AI could wipe out half of all entry-level white-collar jobs within one to five years, specifically targeting professions in law, consulting, and finance. He further cautioned that unemployment could spike to 10–20% and proposed a "token tax" of roughly 3% of AI revenue as a policy cushion. The remarks were tied to rapid capability advances in Anthropic's Claude models and the launch of the company's Economic Index initiative designed to track AI's labor market impact.
The claim arrives amid mounting evidence that generative AI is reshaping professional services hiring. A landmark 2023 OpenAI-affiliated study ("GPTs are GPTs" by Eloundou et al.) estimated that roughly 80% of the U.S. workforce could see at least 10% of their tasks affected by large language models, with approximately 19% of workers facing exposure on 50% or more of their tasks. McKinsey's Global Institute separately projected trillions of dollars in productivity potential from generative AI, noting that document drafting, legal research, data synthesis, and slide generation — staples of junior professional work — are among the most automatable cognitive tasks.
Industry signals lend some weight to the warning. Throughout 2024 and into 2025, major professional services firms including PwC began reducing or restructuring graduate-level hiring, while consulting firms publicly reconsidered the traditional "pyramid" staffing model that relies heavily on entry-level analysts. However, the picture is not uniformly dire: PwC's own 2025 Global AI Jobs Barometer found that AI-exposed occupations saw a wage premium of about 56% in 2024, and the World Economic Forum's 2025 Future of Jobs Report projects a net gain of 78 million jobs globally by 2030 even as 92 million roles are displaced — suggesting significant churn rather than outright elimination.
Fact Check
Evidence from both sides
Supporting Evidence
High task-exposure in white-collar professions
The "GPTs are GPTs" study (Eloundou et al., OpenAI/arXiv, March
found that approximately 19% of U.S. workers could see 50% or more of their tasks affected by LLMs, with legal, finance, and consulting tasks scoring especially high on exposure indices
This supports the plausibility that entry-level roles in these fields are particularly vulnerable.
McKinsey identifies core junior tasks as highly automatable
McKinsey Global Institute's June 2023 analysis concluded that generative AI can rapidly automate research, document drafting, data synthesis, and routine analytical work — functions that constitute a large share of what entry-level lawyers, consultants, and finance professionals do daily. The firm projected trillions of dollars in productivity gains from these specific use cases.
Firms are already cutting entry-level hiring pipelines
Reporting from the Financial Times and consultancy trade press throughout 2024–2025 documented that major professional services firms, including PwC and several top-tier consulting firms, have reduced or restructured graduate-level intakes. Some firms are actively piloting LLM-powered tools to handle tasks previously assigned to junior staff, signaling a structural shift in how the traditional staffing pyramid operates.
Amodei's statement comes from a credible insider position
As CEO of Anthropic — the company behind the Claude family of AI models — Amodei has direct visibility into the pace of capability improvements. His warning was delivered in a formal Axios interview and tied to the launch of Anthropic's Economic Index, a dedicated research initiative tracking AI's economic effects, lending institutional weight to the claim.
Contradicting Evidence
Task exposure does not equal job elimination
The widely cited "GPTs are GPTs" study explicitly measures task-level exposure, not direct job losses. Researchers emphasize that having tasks affected by AI is fundamentally different from having a position eliminated — many roles will be restructured or augmented rather than "completely wiped out," making Amodei's 50% figure far more dramatic than the underlying research supports.
WEF projects net job creation, not mass unemployment
The World Economic Forum's 2025 Future of Jobs Report forecasts approximately 92 million jobs displaced but 170 million created by 2030, yielding a net gain of 78 million positions globally. This suggests large-scale labor market churn and sectoral disruption, but contradicts the narrative of straightforward mass unemployment among white-collar workers.
PwC data shows AI-exposed workers gaining value, not losing jobs
PwC's 2025 Global AI Jobs Barometer, covered by CNBC on June 6, 2025, found that job openings and wages in AI-exposed occupations have actually been growing, with workers possessing AI skills commanding a wage premium of roughly 56% in 2024. This supports an augmentation narrative where AI makes workers more productive and valuable rather than redundant.
Potential conflict of interest in the prediction
Amodei leads a company that directly benefits from widespread belief in AI's transformative — even disruptive — power. Dramatic predictions about job displacement can drive enterprise adoption, government attention, and investment into Anthropic's products. While this does not invalidate his claims, the incentive structure warrants skepticism about the specific magnitude and timeline he offers.
Historical precedent suggests slower displacement timelines
Previous waves of workplace automation — from spreadsheets to robotic process automation — consistently displaced specific tasks faster than they eliminated entire job categories. McKinsey's own research emphasizes task automation and workforce redeployment rather than one-to-one job elimination, suggesting the one-to-five-year window for 50% elimination is likely compressed beyond what historical patterns would predict.
Report an Issue
Found something wrong with this article? Let us know and we'll look into it.