Automation may dismantle the gig economy’s safety net without producing visible layoffs

FT columnist Sarah O’Connor argues that automation is beginning to erode both halves of the gig economy. Generative AI is taking simple copywriting and design jobs from platforms such as Fiverr and Upwork, which are trying to move toward higher-value services. Physical automation is approaching ride-hailing and delivery work as robotaxis expand in the US and China; Uber has lobbied for a slower rollout even while partnering with autonomous-vehicle companies.

The underappreciated policy problem is measurement. Gig workers are not formally laid off when an algorithm takes demand. They simply receive fewer tasks, wait longer or accept lower rates. In China, delivery and ride-hailing have absorbed millions of people excluded from formal employment; a gradual reduction in available work could weaken that safety valve without appearing in conventional redundancy statistics.

New platform work in healthcare and AI training may grow, but it requires different skills and is unlikely to absorb everyone displaced from driving or commodity creative work. The accompanying Rest of World cases show the same transition from the worker’s side: AI can let one person perform more functions, while simultaneously reducing the number of specialists a low-budget project hires.

Governments built little social protection around gig work because it was framed as flexible entrepreneurship; automation could now remove that informal safety net in a way that is economically large but statistically quiet.