Amazon wired human judgment into its catalog cleanup

As Amazon's catalog grew far beyond books, it accumulated large numbers of duplicate product pages — the same item listed multiple times with slightly different titles, images, or descriptions. Engineers tried automating the cleanup with matching algorithms, but the task kept defeating software: a human could glance at two listings and tell in seconds whether they were the same product, while writing rules general enough to catch every variation without producing false matches proved impractical at catalog scale.

Amazon manager Venky Harinarayan's team built what he called a hybrid machine-human computing arrangement: a system that broke the deduplication task into small, independent units, each answerable by a human in seconds, and distributed those units automatically over the internet to a pool of remote workers who could complete them whenever and wherever they chose. Jeff Bezos saw the internal tool's broader potential and had it turned into a public marketplace in 2005, named Mechanical Turk after the 18th-century chess-playing automaton that had secretly concealed a human operator inside.

Opened a public marketplace for micro-tasks like duplicate detection, paid per task, so software could call human judgment as an API and scale labor with demand.

Microtasks need no context, so any worker can answer quickly.

Distributed workers scale with demand, not headcount.

Programmatic calls make human judgment as easy to use as software.

When automation fails at a task needing quick human judgment, decompose it into microtasks and route to on-demand humans via an API.

The internal tool solved Amazon's own catalog problem, then launched publicly in 2005 as a marketplace still running today. Mechanical Turk became the template for the broader 'crowdwork' and human-in-the-loop AI industry, later underpinning much of the labeled training data behind modern machine learning systems, all descended from a tool built to solve a strictly internal catalog problem.