AI Will Transform Capitalism, but Who Will Shape Its Future?

Summary

Paul Mason argues that artificial intelligence could do more than automate workers’ tasks: by replacing parts of entrepreneurial, innovative and financial work, it could weaken the connection between human labor, innovation and profit that supports capitalism. He places this possibility in a longer intellectual tradition, from Aristotle’s imagined self-operating tools to Marx’s “general intellect,” while noting that current large language models already absorb broad bodies of human knowledge and can exceed the output of some experienced knowledge workers. Mason contrasts two competing geopolitical strategies. The US approach, as he describes it, concentrates computing power and offers access through online services, encouraging dependence on American software, cloud infrastructure and rules. China’s counterstrategy emphasizes open-source models that can be run locally, allowing users to keep data behind their own firewalls while supporting Beijing’s bid for technological leadership in fields built on AI. He rejects the idea that Chinese models such as Kimi, Qwen and DeepSeek automatically signal the end of capitalism, calling them an alternative route to technological dominance. The essay presents a third possibility: shaping AI around human need through universal basic services and incomes, cooperative or nonprofit ownership, and AI-assisted economic planning. Mason proposes developing an unownable model guided by humanistic, universal and environmentally conscious goals, even if it trails leading commercial systems. His conclusion is an argument rather than a prediction: AI does not automatically produce a classless society, but its disruption of employment, ownership and profit could create an opening for alternative social arrangements.