Amazon’s Mechanical Turk is not being switched off, but its future has narrowed sharply. On July 30, 2026, the crowdsourcing marketplace will close to new customers, while existing customers will be allowed to keep using it as normal.
A Long-Running Crowdsourcing Service Enters a Smaller Phase
An announcement on the Mechanical Turk website says the service will stop taking new customers on July 30, 2026. Amazon Web Services says the decision came after “careful consideration.”
AWS also said, “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.”
That statement draws a clear line around the product. Mechanical Turk is staying available for customers already inside the system, but the service is no longer being positioned for growth. The lack of planned new features matters because marketplaces depend on both supply and demand: people who post tasks, and people who complete them.
Closing the door to new customers does not end that marketplace overnight. But it does change the signal from Amazon. A platform that once sat at the center of conversations about online labor, automation and AI is now being maintained rather than expanded.
What Mechanical Turk Was Built to Do
Mechanical Turk first launched in 2005. Its basic idea was simple: pay people small amounts to complete tasks that computers could not easily handle on their own.
Those tasks included things like completing CAPTCHA challenges or identifying the basic sentiment in a sentence. The work was often simple at the level of each individual task, but useful at scale because many small human judgments could be gathered through the same marketplace.
The platform became known as a way to put human judgment behind digital systems. It also became a focal point for debates around the ethics of crowdsourced labor. The core tension was easy to see: companies could route work to a large online labor pool, while workers were often paid tiny amounts for small tasks.
Mechanical Turk also had a visible place in broader technology history. The service played a small role in the early stages of the Facebook-Cambridge Analytica scandal, according to the source article.
AI Made the Mechanical Turk Story More Complicated
Beginning in 2018, Amazon also began billing Mechanical Turk as a way for companies to annotate data to train neural networks as part of its SageMaker AI service. That tied the platform more directly to the rise of AI development.
Data annotation depends on labeling, sorting or evaluating information so that systems can learn from it. Based on the source article, Mechanical Turk was presented as one path for companies that needed that kind of human input.
At the same time, Mechanical Turk had a less direct but important connection to AI. It has been described as the hidden enabler for companies taking a fake-it-till-you-make-it approach to AI, where products marketed as Ai are actually being performed by the Mechanical Turk workforce.
That connection is especially pointed because the original Mechanical Turk was itself a hoax. It involved a hidden human chess player pretending to be a chess-playing machine. The modern service carried the same name into a digital era where the line between human work and machine intelligence kept getting harder to see.
When Workers Use AI to Do Human Tasks
The relationship between Mechanical Turk and AI did not stay simple. A 2023 analysis found that between 33% and 46% of workers on the platform were using large language models to complete their tasks.
That finding raised two related questions. First, if a company is paying for human-labeled data, what happens when some of that work is actually produced with large language models? Second, if AI systems can perform a meaningful share of the tasks, how much human involvement is still needed?
Those questions are especially important for data annotation. If the goal is to train or evaluate AI systems using human judgment, the reliability of the work depends on knowing what kind of judgment is actually being supplied.
The 2023 analysis did not just complicate the value of the platform. It also pointed to a broader issue for online labor markets built around small, repeatable tasks. When automation becomes good enough to complete those tasks, the platform can become both a tool for AI development and a place where AI substitutes for the human work it was supposed to organize.
What Happens Next for Mechanical Turk
Amazon has not said it is fully ending Mechanical Turk. Existing customers can continue using the service as normal, and AWS says it continues to invest in security and availability improvements.
But the decision to block new customers and stop planning new features leaves the platform in a constrained position. It can continue to operate, yet it no longer appears to be a product Amazon is trying to expand.
After the decision became public, one Reddit user suggested the platform died “years ago,” saying workers and researchers had moved away from it because of bots and fraud. The same user predicted, “Someone at Amazon is going to decide keeping the Mturk servers running is a waste of time and resources and pull the plug entirely.”
That prediction is not an Amazon announcement. But it captures the uncertainty now surrounding Mechanical Turk. A service that began in 2005 as a marketplace for small human tasks became part of the AI data pipeline, then became entangled with questions about whether those tasks were still being done by people at all.
For now, Mechanical Turk remains available to existing customers. The larger story is that one of the better-known names in crowdsourced labor is no longer accepting newcomers, and its next chapter looks more limited than its past.