A new report from McKinsey finds that currently available AI technologies could, in principle, automate tasks equivalent to about 57% of today’s U.S. work hours — mostly non-physical, knowledge-based tasks handled by digital “agents.” Robots and other physical automation could cover another 13% of work hours, primarily tasks involving manual labor.
However, the report emphasises this doesn’t mean half of all jobs will disappear. Rather, what’s likely to happen is a shifting division of labor: many roles will transform as routine or repetitive tasks get automated, while humans take on new responsibilities. In this shift, humans are expected to supervise, validate, and guide AI and automation — especially when tasks call for judgment, empathy, or social-emotional intelligence.
Crucially, the study argues that tasks requiring real-time perception, empathy, ethical judgment, creativity, or complex interpersonal interaction remain beyond what AI can reliably handle — at least for now. Roles such as teaching, nursing, counselling, sales with nuanced human feedback, or customer-facing functions likely remain rooted in human effort, because those jobs rely heavily on social and emotional intelligence.
Looking ahead, the report envisions a future workforce that isn’t one of wholesale replacement, but of partnerships — between people, AI-powered agents, and robots. As automation expands, some traditional tasks will vanish or shrink, others will evolve, and novel job categories will emerge — especially around AI supervision, orchestration, quality control, and hybrid human-machine workflows.
