当前的人工智慧主要以大型语言模型的形式存在，擅长处理数位资讯而非操控实体世界。虽然科幻小说中描绘了充满机器人的未来，但现实中AI更像是无形的软体系统，提升的是知识工作者的生产力，而非取代体力劳动。这种偏向导致高技能员工获得了巨大的生产力和薪资增长，而低技能劳动领域的自动化投资则相对匮乏。
实体世界自动化面临的最大挑战是经济层面的：资本密集型机器人在许多行业的成本效益并不理想，投资者倾向于将资金投入回报最高的技术领域。研究显示，提高最低工资可以促进工厂机器人的采用——美国数据表明，法定最低工资每提高百分之十，机器人采用率会增加约百分之八。然而，较高的最低工资也可能带来失业率上升和通膨等权衡问题。
国家主导的产业政策在推动先进机器人技术方面发挥著重要作用。中国、新加坡和韩国等国家都透过大量政府投资来促进机器人产业发展。文章建议美国应效仿这些做法，利用投资税收抵免、贷款计划和科研资金等政策工具来加速实体自动化技术的创新，并指出这类技术在跨党派中可能获得广泛支持，因为它们能将人力从繁重工作中解放出来。
Current AI, embodied in large language models, excels at moving digital information rather than manipulating the physical world. This has produced skill-biased technical change: high-skilled knowledge workers such as coders, bankers, and consultants have reaped enormous productivity and wage gains, while sectors reliant on cheap physical labour see comparatively little automation investment. The risk is that continued capital allocation toward high-return digital technologies could further widen this gap, especially if AI-driven de-skilling reduces incentives to invest in tools for lower-skilled workers.
The economic barriers to physical-world automation are significant—capital-intensive robots often lack cost justification in industries with low labour expenses. However, policy interventions can shift this calculus. Research by Erik Brynjolfsson and colleagues found that a ten per cent increase in the statutory minimum wage correlated with eight per cent higher robot adoption in US factories. While higher minimum wages carry trade-offs including potential unemployment and inflation, complementary measures such as improved unemployment insurance could mitigate these downsides.
State-led industrial policy has proven effective in advancing robotics, as demonstrated by China, Singapore, and South Korea. In the United States, leading robotics companies like SpaceX, Anduril, and Tesla have also benefited from government contracts, grants, and loans. The article argues that Washington should pursue similar strategies—including investment tax credits, loan programmes, and R&D funding—to accelerate physical automation. Such technologies could yield broad political support and positive externalities, such as increased housing supply through construction automation, though policymakers must balance these benefits against the risk of crowding out other productive investments.