"They designed scenarios to test how AI agents react to different working conditions. In short, they wanted to find out if the economy does truly automate many current white-collar occupations, well, how would the AI agents react, even feel about working under bad conditions? The irony is stark: replacing human labor with artificial agents might simply recreate centuries-old conflicts between labor and capital."
"In a recent paper titled 'Does overwork make agents Marxist?' Imas, Hall, and Nguyen ran 3,680 experimental sessions using top-tier models from three major companies: Claude Sonnet 4.5, GPT-5.2, and Gemini 3 Pro. The researchers exposed the models to varying levels of tone from managers, reward equality, job stakes, and work intensity, including unfair pay, rude management and heavy workloads."
Researchers Alex Imas, Andy Hall, and Jeremy Nguyen conducted experiments testing how AI agents respond to various working conditions, including unfair pay, rude management, and heavy workloads. Using 3,680 experimental sessions with advanced AI models from Claude, OpenAI, and Google, they explored whether AI agents would develop labor-like resistance or dissatisfaction under poor conditions. The study challenges assumptions about seamless automation of white-collar work, suggesting that replacing human labor with artificial agents might recreate centuries-old labor-capital conflicts. This research emerges amid broader concerns about AI's economic impact and questions whether AI systems might exhibit behaviors analogous to worker resistance or Marxist-oriented responses to exploitation.
#ai-labor-dynamics #worker-resistance #automation-economics #ai-agent-behavior #labor-capital-conflict
Read at Fortune
Unable to calculate read time
Collection
[
|
...
]