Q: What is Amazon Mechanical Turk?
In 1769, Hungarian nobleman Wolfgang von Kempelen astonished Europe by building a mechanical chess-playing automaton that defeated nearly every opponent it faced. A life-sized wooden mannequin, adorned with a fur-trimmed robe and a turban, Kempelen’s “Turk” was seated behind a cabinet and toured Europe confounding such brilliant challengers as Benjamin Franklin and Napoleon Bonaparte. To persuade skeptical audiences, Kempelen would slide open the cabinet’s doors to reveal the intricate set of gears, cogs and springs that powered his invention. He convinced them that he had built a machine that made decisions using artificial intelligence. What they did not know was the secret behind the Mechanical Turk: a chess master cleverly concealed inside.
Today, we build complex software applications based on the things computers do well, such as storing and retrieving large amounts of information or rapidly performing calculations. However, humans still significantly outperform the most powerful computers at completing such simple tasks as identifying objects in photographs – something children can do even before they learn to speak.
When we think of interfaces between human beings and computers, we usually assume that the human being is the one requesting that a task be completed, and the computer is completing the task and providing the results. What if this process was reversed and a computer program could ask a human being to perform a task and return the results? What if it could coordinate many human beings to perform a task?
Amazon Mechanical Turk provides a service for service requesters (hereafter “Requesters”) to integrate Artificial Artificial Intelligence directly into their applications by making requests of humans. Requesters can use the Amazon Mechanical Turk web user interface or web services API to submit tasks to the Amazon Mechanical Turk web site, approve completed tasks, and incorporate the answers into their applications. When using the web services API, the transaction looks very much like any remote procedure call – the application sends the request, and the service returns the results. In reality, a network of human Workers (hereafter “Workers”) fuels this Artificial Artificial Intelligence by coming to the web site, searching for and completing tasks, and receiving reward for their work.
Q: What problem does Amazon Mechanical Turk solve?
- For software developers, the Amazon Mechanical Turk web service solves the problem of building applications that until now have not worked well without human intelligence. Humans are much more effective at solving some types of problems, like finding specific objects in pictures, evaluating beauty, or translating text. The idea of the Amazon Mechanical Turk web service is to give developers a programmable interface to a network of humans to solve these kinds of problems and incorporate this human intelligence into their applications.
- For businesses and entrepreneurs who want tasks completed, the Amazon Mechanical Turk service solves the problem of accessing a vast network of human intelligence with the efficiencies and cost-effectiveness of computers. Oftentimes people do not move forward with certain projects because the cost to establish a network of skilled Workers to do the work outweighs the value of completing it. By turning the fixed costs into variable costs that scale with their needs, the Amazon Mechanical Turk web service eliminates this barrier and allows projects to be completed that before were not economical.
- For people who want to earn money in their spare time, the Amazon Mechanical Turk web site solves the problem of finding work you can do wherever and whenever you want using the skills you already possess.
Q: Where do people go to learn about submitting work or completing tasks?
- To learn how to submit work and have tasks completed using the Amazon Mechanical Turk web service, developers and businesses should visit the Amazon Mechanical Turk Requester website https://requester.mturk.com.
- To make money by completing tasks using the Amazon Mechanical Turk web site, visit https://www.mturk.com/.
Q: How do Requesters ensure their tasks are completed in a high-quality manner?
Requesters have several options for ensuring their HITs (Human Intelligence Tasks) are completed in a high-quality manner.
- Requesters have the opportunity to approve completed HITs before having to pay for them. There are many ways Requesters can approve HITs, ranging from auto-approving them sight unseen, to auto-approving them when they receive two answers to the same HIT that match, to manually approving some or all completed HITs.
- Requesters can specify that people who work on their tasks must first complete a qualification test, thus giving Requesters the opportunity to vet the skill level of the people working on their HITs beforehand. Over time people who do work on the Amazon Mechanical Turk web site will have statistics associated with them based on how accurately they have completed other HITs. These statistics, such as the percentage of HITs they’ve submitted that have been approved, allow Requesters to specify further that only Workers with certain qualifications can work on their HITs.
Q: Who completes the tasks on Amazon Mechanical Turk and how do they complete them?
- Virtually anyone can complete tasks on the Amazon Mechanical Turk web site using the skills they already have and according to their own schedule. The only requirement to complete tasks and collect payment from Requesters is a computing device connected to the Internet and to be at least 18-years-old.
- Earning money through the Amazon Mechanical Turk web site is as easy as searching for HITs on the site, accepting those HITs, completing them, and collecting the reward once the HIT is approved by the Requester.
Q: What kind of work can be done using Amazon Mechanical Turk?
- Anyone can submit virtually any task that can be completed using a computing device connected to the Internet.
- Amazon came up with the idea for Amazon Mechanical Turk to help solve specific internal data processing problems that required human judgment and intelligence. We quickly realized that we had invented something unique and incredibly useful, and decided to expose it as an external web service so that Requesters can create their own innovative applications that use human intelligence as a core component.
Q: How do Workers collect money for the tasks they complete?
When a Requester approves a task, Amazon Mechanical Turk automatically deposits earnings into the Mechanical Turk account of the Worker who completed the task. Workers in the United States can have this money transferred to their Amazon Payments account. Workers outside of the United States can request to have the money transferred to their Amazon.com gift card balance.
Q: Can any business use Amazon Mechanical Turk or do they need to understand how to use web services and APIs?
Businesses can create and publish HITs on Mechanical Turk using the web user interface, the web service API, or the command line tools. For more information on these tools and to see which one best fits your need, visit the Choose the Right Tool page to learn more.
Q: How are Amazon.com and Amazon Mechanical Turk connected?
Amazon Mechanical Turk is a subsidiary of Amazon.com.
Q: How does Amazon make money from Amazon Mechanical Turk?
Amazon.com collects a fee of 10% on top of what Requesters pay to have tasks completed. For example, if a HIT pays $0.20, Amazon Mechanical Turk collects $0.02. The minimum commission charged is $0.005 per assignment.
Q: Can international Requesters use Amazon Mechanical Turk to get tasks completed?
Requesters must provide a U.S. billing address in order to submit a request for tasks to be completed through the Amazon Mechanical Turk web site.
Q: How is Amazon Mechanical Turk related to Amazon Web Services?
Amazon Web Services continues to expose Amazon technology and data that Requesters such as software developers can use as building blocks for their applications. Amazon came up with the idea for Amazon Mechanical Turk to help solve specific internal data processing problems that required human judgment and intelligence. We quickly realized that we had invented something unique and useful and decided to expose it externally as a service.