Amazon Winds Down Mechanical Turk, the Web's Original Gig Engine

Amazon is quietly retiring one of the internet's most influential labour experiments. The crowdsourcing marketplace that helped train a generation of machine-learning systems is being wound down, just as the technology it fed reaches the mainstream.

The News

Amazon Web Services will stop accepting new customers for Mechanical Turk from 30 July 2026, according to updated AWS documentation. Businesses already using the platform can carry on, but the company has been blunt about its ambitions: it does not plan to introduce new features, and the decision was taken after what it called careful consideration.

Launched in 2005, Mechanical Turk let companies break large jobs into thousands of micro-tasks and farm them out to a distributed pool of online workers for a few cents each. Sentiment tagging, image labelling, transcription and CAPTCHA-style verification were the staples. AWS said it continues to invest in security and availability, but the freeze on new sign-ups marks the beginning of the end for the two-decade-old service.

The timing is pointed. A 2023 study found that between 33% and 46% of Mechanical Turk workers were themselves using large language models to complete assignments, muddying the very human-labelled datasets the platform was supposed to guarantee.

Why It Matters

Mechanical Turk was, in effect, the scaffolding beneath the current AI boom. Long before foundation models could label their own training data, human annotators on platforms like this one produced the tagged images, ranked answers and moderation judgements that taught algorithms to see, read and rank. By 2018 Amazon had repositioned the tool explicitly as a data-annotation engine inside its SageMaker machine-learning stack.

Its decline mirrors a broader shift. When the last comparable inflection arrived, the release of ChatGPT in late 2022, the assumption was that cheap human microwork would scale forever alongside AI. Instead, the opposite has happened: models have grown capable enough to automate the simplest labelling tasks, while the hardest work has moved upmarket to specialist firms employing trained, better-paid annotators. A marketplace built for anonymous penny-per-task labour no longer fits either end of that barbell.

The platform also carried reputational baggage, from long-running criticism over low wages to its peripheral role in the Facebook-Cambridge Analytica episode. Retiring it lets Amazon exit a low-margin, ethically fraught business while pouring resources into higher-value AI infrastructure.

Indian Angle

Few markets should read this news more closely than India. For years India and the United States supplied the largest shares of Mechanical Turk's global workforce, with students, homemakers and small-town graduates treating micro-tasking as flexible supplementary income. The wind-down removes a modest but real earnings channel for thousands of Indian gig workers.

More strategically, it validates the model that Indian data-annotation startups have been building. Firms such as iMerit, Karya and the Bengaluru-born Playment (later acquired by Canada's Telus) bet early that enterprises would pay a premium for managed, quality-controlled human labelling rather than an open marketplace of anonymous contractors. Amazon's retreat from raw crowdwork is a signal that the future belongs to exactly this curated, accountable approach, where India already holds an edge in cost and English-language capacity.

There is a policy dimension too. As MeitY frames India's data-labelling ambitions and the country positions itself as a global AI services hub, the fading of an American platform leaves room for domestic players to capture enterprise contracts, provided they can guarantee the auditable, LLM-free datasets that buyers now demand.

FAQ

When does the change take effect?

Amazon stops accepting new Mechanical Turk customers on 30 July 2026. Existing customers can keep using the service after that date, but Amazon has confirmed it will not add new features and is effectively placing the platform into maintenance mode.

Why is Amazon closing it to new customers?

Amazon cited careful consideration without giving a single reason. Analysts point to falling demand as AI automates simple tasks, quality problems from workers using chatbots, thin margins and long-standing criticism over labour conditions on the platform.

What does it mean for Indian data-labelling firms?

It strengthens the case for managed annotation providers such as iMerit and Karya. Enterprises seeking verified, high-quality human-labelled data may shift towards these curated Indian vendors rather than open crowdwork marketplaces, potentially expanding contract opportunities for the sector.

Where can I read the original announcement?

The closure was reported by TechCrunch, drawing on updated AWS documentation. The full coverage is linked in the attribution below.

This story was reported by TechCrunch. Read the full original coverage at TechCrunch.