IT & AI HistoryA Chronicle of AI and Information Technology

Detailed · 10 events

AI and Jobs: A History

The question of AI taking jobs began in earnest with a 2013 Oxford estimate that 47% of US employment was at high risk. In 2015 Nomura Research Institute put 49% of Japan's workforce as replaceable; after generative AI came an investment bank's figure of 300 million jobs exposed, hiring pauses and cuts at large firms, and the writers' and actors' strikes. Who produced each estimate and how, what critics said later, and what the employment data actually showed.

2010s

The jobs debate of the 2010s began with probabilities assigned to whole occupations. The Oxford authors labelled a sample of occupations by hand and let a classifier score the rest; Nomura Research Institute ran the same method over Japanese occupational data and found nearly half the workforce technically replaceable. The Oxford paper did not claim to count jobs that would really disappear, and NRI said it left out social factors and the replacement of only part of a job. In 2016 an OECD paper, measuring tasks instead of occupations, found a far smaller automatable share. Estimates in the generative-AI years would be rebuilt from task data.

  1. EVT.001T1Frey and Osborne's "The Future of Employment" — The Paper That Put 47% of US Jobs at High RiskAI and Society: A HistoryA History of Artificial IntelligenceRead more →
  2. EVT.002T1Nomura Research Institute's "49% of Japan's Workforce Replaceable" — Where Japan's AI-and-Jobs Debate BeganAI and Society: A HistoryA History of Artificial IntelligenceRead more →

2020s

In the 2020s the jobs question moved from forecasts to company decisions and payroll data. The big tech cuts of early 2023 were explained as hiring for a pandemic economy that had passed. Once generative AI spread, Goldman Sachs extrapolated to 300 million full-time jobs exposed, and executives at IBM and Klarna described how much work AI could take on. Later accounts wavered: Klarna's chief admitted quality had suffered, and Amazon's said its cuts were about culture rather than AI. Stanford researchers found young workers in exposed occupations losing ground, while Yale's Budget Lab saw no broad disruption.

  1. EVT.003T1The 2023 Tech Layoffs — The Month a Decade of Hiring EndedA General History of Information TechnologyRead more →
  2. EVT.004T1Goldman Sachs's "300 Million Jobs" Estimate — What "Exposed" to Generative AI MeantAI and Society: A HistoryA History of Artificial IntelligenceRead more →
  3. EVT.005T1IBM's "7,800 Jobs" AI Hiring Pause — When a Big-Company CEO Put a Number on AI ReplacementAI and Society: A HistoryA History of Artificial IntelligenceRead more →
  4. EVT.006T1Hollywood's Double Strike — Writers and Actors Write AI into the ContractAI and Society: A HistoryA History of Artificial IntelligenceAI and Copyright: A HistoryRead more →
  5. EVT.007T1Klarna's AI Assistant and the "Work of 700 Agents" — Replacing Customer Service, Then Bringing People BackAI and Society: A HistoryA History of Artificial IntelligenceRead more →
  6. EVT.008T1Devin — The Agent Billed as 'the First AI Software Engineer'A History of Artificial IntelligenceAI and Society: A HistoryA History of AI AgentsRead more →
  7. EVT.009T1"Canaries in the Coal Mine" — Evidence That AI Is Cutting Entry-Level Jobs, and the Case AgainstAI and Society: A HistoryA History of Artificial IntelligenceRead more →
  8. EVT.010T1Amazon's 14,000 Corporate Job Cuts — A CEO Who Wrote "AI Will Reduce Our Workforce", and an Explanation Called "Culture"AI and Society: A HistoryA History of Artificial IntelligenceRead more →