Amazon Closes Mechanical Turk Crowdsourcing Platform September 30
Amazon will permanently close Mechanical Turk on September 30, 2026, ending its long-running marketplace for human data-labeling, transcription, survey, and review tasks. AWS documentation advises SageMaker Ground Truth and Amazon Augmented AI users to transition affected workflows to another workforce option before the closure. The shutdown ends a long-running source of on-demand human labor as AI models advance and new data-labeling providers enter the market.
Amazon will permanently close its Mechanical Turk crowdsourcing platform on September 30, 2026. A closure notice quoted by CNBC states that Amazon made the decision after evaluating its programs, tools, and services, while AWS documentation instructs customers using the MTurk workforce to review affected workflows and transition to an alternative before that date.
Launched in 2005, Mechanical Turk, commonly called MTurk, connected requesters with workers for small digital jobs known as Human Intelligence Tasks. CNBC and Quartz report that common uses included data labeling, audio and video transcription, and surveys, with workers generally paid cents per task. Amazon founder Jeff Bezos previously described the service as "artificial artificial intelligence," according to CNBC, referring to work that humans could complete more reliably than computers.
Implications for AWS labeling workflows
The closure directly affects users of the public MTurk workforce in Amazon SageMaker Ground Truth labeling jobs and Amazon Augmented AI, or A2I, human-review workflows. AWS documentation describes MTurk as a worldwide, around-the-clock workforce that typically offered fast turnaround for labeling and review tasks.
AWS documentation also notes important data-handling limits: Ground Truth and A2I inputs used with the MTurk workforce had to be free of personally identifiable information, and the workforce could not be used for protected health information in conjunction with AWS HIPAA-eligible services. Those restrictions meant MTurk was not a universal human-in-the-loop option, but its removal still requires affected teams to change the workforce component of existing pipelines.
Amazon's documentation does not name a replacement provider or a migration path beyond moving to an alternative workforce option. It also does not cite AI as the reason for the shutdown. The public notice quoted by CNBC gives the company's stated rationale only as an assessment of its programs, tools, and services.
A changing market for human data work
Quartz reports that MTurk once served more than 500,000 workers. CNBC quotes Krista Pawloski, a data worker and organizer with Turkopticon, who characterized the platform as having been in decline in recent years. Quartz reports that Pawloski attributed part of that decline to less Amazon investment in the service and competition from data-labeling providers.
Business Insider reports that Amazon stopped accepting new MTurk users last month. Quartz identifies Scale AI, Mercor, and Prolific among newer providers recruiting workers for AI-training-related work.
The closure arrives as model development has changed the economics and requirements of data work. CNBC reports that AI models have advanced substantially since MTurk launched, and Business Insider notes that some tasks once routed through the marketplace can now be automated. Separately, companies undertaking comparable shifts from open crowdsourcing to managed data vendors often trade broad labor availability for more structured worker qualification, quality controls, and data-governance processes.
For ML teams, the immediate operational question is narrower than the broader labor-market narrative: identify Ground Truth and A2I workflows using the public MTurk workteam ARN, validate replacement workforce capacity and quality, and complete the transition before September 30. The broader market question is whether generalized task marketplaces can remain viable where automation handles routine work and remaining annotation requires more specialized expertise.
Key Points
- 1Amazon will close Mechanical Turk on September 30, requiring SageMaker Ground Truth and A2I users to replace its public workforce.
- 2MTurk supported labeling and human review, making workforce migration a practical dependency for affected ML data pipelines.
- 3Newer providers such as Scale AI, Mercor, and Prolific have entered the market for AI-training-related work.
Scoring Rationale
The shutdown affects a widely known human-data-work platform and creates a concrete migration deadline for AWS customers using its SageMaker labeling and A2I review integrations. It is not a new model or infrastructure release, but it is notable business and workflow news for teams that depend on external human annotation.
Sources
Primary source and supporting public references used for this report.
Practice with real Retail & eCommerce data
90 SQL & Python problems · 15 industry datasets
250 free problems · No credit card
See all Retail & eCommerce problems
