AI Agents won’t just peplace jobs, they’ll reshape human value at work
A Stanford study by Humishka Zope and colleagues introduces the first large-scale audit of how AI agents align with what workers actually want automated.
By surveying 1,500 workers across 104 occupations, the researchers built the WORKBank database and created the Human Agency Scale (HAS) to measure where humans want collaboration versus full automation.
The results:
– Workers welcome automation for repetitive, low-value tasks but demand human agency in complex, interpersonal, or uncertain work.
– There’s a growing desire–capability mismatch: AI can already automate far more than workers feel comfortable handing over.
– Core skills are shifting. As information-heavy tasks are automated, human strengths in empathy, judgment, and collaboration are becoming the new workplace currency.
This is not just about efficiency. It’s a roadmap for how AI and humans can coexist and a warning that reskilling must start now to protect agency, not just jobs.
