Amazon will stop accepting new customers for Mechanical Turk
The Sunset of Amazon Mechanical Turk
The era of Amazon’s crowdsourcing experiment is drawing to a close. According to a recent announcement on the platform's official site, Mechanical Turk (MTurk) will cease accepting new customer registrations starting July 30, 2026.
The Current Status of the Service
While the doors are closing for newcomers, Amazon Web Services (AWS) has clarified the roadmap for those already onboard. The service is effectively moving from active growth to maintenance mode.
"Existing customers can continue to use the service as normal. AWS continues to invest in security and availability improvements for Mechanical Turk, but we do not plan to introduce new features."
What this means for the platform:
- Maintain server stability and security.
- Support current active accounts.
- Develop new tools or functionalities.
- Onboard new business clients.
A Brief History of Human Intelligence Tasks (HITs)
Launched back in 2005, Mechanical Turk functioned as a digital marketplace for "micro-tasks"—small jobs that were too simple for complex software but too tedious for high-paid employees.
| Era | Primary Focus | Key Activities |
|---|---|---|
| Early Days (2005+) | General Crowdsourcing | Solving CAPTCHAs, sentiment analysis |
| Growth Phase | Ethical Debate | Labor rights discussions, Cambridge Analytica role |
| AI Pivot (2018+) | Data Labeling | Annotating datasets for SageMaker AI |
The Great AI Irony
There is a poetic, if cynical, symmetry to the platform's decline. The original "Mechanical Turk" of the 18th century was a famous hoax—a machine that appeared to play chess but actually hid a human inside. In a similar vein, MTurk became the "hidden engine" for many modern AI startups. Many companies marketed "AI-powered" products that were actually just humans on MTurk performing tasks behind a digital curtain.
However, the relationship eventually became a "snake eating its own tail." A 2023 study revealed a startling trend in the workforce:
This means nearly half of the "human" workers were using Large Language Models to complete the tasks meant to train those very models. This creates a feedback loop that threatens the integrity of the data.
The Cycle of Automation
The Final Verdict
The community sentiment suggests this move was inevitable. On Reddit, users have argued that the platform's utility vanished years ago, eroded by a plague of bots and fraudulent activity.
One user summarized the likely endgame in a bleak prediction:
# Predicted outcome for MTurk servers
if (cost_of_maintenance > perceived_value) {
amazon_action = "pull_the_plug_entirely";
status = "OFFLINE";
}
About the Author:
Anthony Ha is a seasoned journalist with a background at VentureBeat, Adweek, and various VC firms. He can be reached via email at anthony.ha@techcrunch.com.
Related Events:
Don't miss the TechCrunch Founder Summit in Boston on November 4th!