The scenes that Metz sees don’t look much like the future we imagine, or at least the automation future you might imagine. The office can be a call center or payment processing center, one of which is located in an old apartment building in the middle of the low-income residential area in western Kolkata, crowded with pedestrians, tricycles and street vendors. In Bhubaneshwar, where he visited, and other cities in India, Nepal, the Philippines, East Africa and the United States, tens of thousands of office workers are working on training machines.

There are tens of thousands of workers, independent contractors who usually work from home, and also annotate data through crowdsourcing services such as Amazon Mechanical Turk, a service that allows anyone to assign digital tasks to the US and Independent workers in other countries, workers can earn a few cents per label.

IMerit, headquartered in India, labels data tags for many of the big names in the technology and automotive industries. The company refused to disclose the names of these customers on the grounds of a confidentiality agreement. But the company recently revealed that its more than 2,000 employees in nine offices around the world are contributing to Amazon’s online data labeling service, SageMaker Ground Truth. Previously, it also listed Microsoft as a customer.

Uncover the secrets of AI training: In addition to experts and help the AI ​​evolution, there are countless office workers around the world

▲ Figure 3: Artwork presented at the iMerit office in the Metiabruz community in Kolkata, India

To be sure, AI may short open the job market in the future. But for now, it is creating jobs with relatively low incomes. According to research firm Cognilytica, the market value of data tags in 2018 is more than $500 million and will reach $1.2 billion by 2023. Studies have shown that this type of work accounts for 80% of the time spent building AI technology.

Is this work exploitative? It depends on where you live and what you are doing. In India, this is a ticket to the middle class. This is a decent job in New Orleans, USA. But for those who are independent contractors, this is often a “no return”.

Some skills must be learned, such as in video or medicineThe company recently opened a new office in Metiabruz, a Muslim-dominated community in western Calcutta. There, it employs mostly Muslim women whose families are reluctant to let them leave this bustling area. They are not required to see erotic images or violent materials.

Understanding AI Training Insider: In addition to experts who help AI evolve, there are countless office workers around the world

▲Figure 7: iMerit employees receive training at the office of Metiabruz in Kolkata

At first, iMerit focused on simple tasks, sorting out product listings for online retail sites, and reviewing posts on social media, but it has moved to work that supports AI. The growth of iMerit and similar companies represents a shift from crowdsourcing services like Mechanical Turk. iMerit and its customers have greater control over how employees are trained and how they work.

Baidia is now the manager of iMerit, who is responsible for tagging the street scenes used by a large American company to train driverless cars. His team analyzed and tagged digital photos and three-dimensional images captured by lidar. They draw border boxes around cars, pedestrians, parking signs and wires all day long.

Baidia said the job might be boring, but it gave him a life he might not have. He and his wife recently bought an apartment in Calcutta and walked to the iMerit office where she works. Baidia said: “My life has undergone a fantastic change, both in terms of my financial situation, personal experience and English skills. I got an opportunity!”

Listen to people coughing

Under AI training insider: In addition to experts in the AI ​​evolution, there are countless office workers around the world

▲ Figure 8: Oscar Cabezas working at the iMerit New Orleans office. When the company started developing Spanish digital assistants, he joined the company

A few weeks after the trip to India, Metz took Uber through downtown New Orleans. About 18 months ago, iMerit moved into a building across Superdome Street. A large US technology company needs a way to tag data for the Spanish version of its home digital assistant. Therefore, it sends the data to the new iMerit office in New Orleans.

After Hurricane Katrina in 2005, hundreds of construction workers and their families moved to New Orleans to help rebuild the city, and many stayed. Many Spanish-speaking people came with this new workforce and the company began to hire them.

23-year-old Oscar Cabezas and her mother moved from Colombia to New Orleans. His stepfather found a job on the construction site. After graduating from college, Kabesas joined iMerit and began developing Spanish digital assistants.

He commented on everything from tweets to restaurant reviews, identifying people and places, and finding ambiguities. For example, in Guatemala, “pisto” means money, but in Mexico it means beer. He said: “There are new projects every day.”

This office has expanded into other areas to serve businesses that want to keep their data in the United States. Some projects must remain in the United States for legal and security purposes.

42-year-old Glenda Hernandez was born in Guatemala and said she missed her previous work on the digital assistant project. She likes to read books, has commented on books online for large publishing companies, so she can get a free copy, she enjoys the paid reading opportunities brought by reading in Spanish.

Understanding AI Training Insider: In addition to experts who help AI evolve, there are countless office workers around the world

Figure 9: Glenda Hernandez is a member of iMerit in New Orleans who has learned to distinguish the difference between cough and good

Hernandes marks images or similarProjects that annotate people’s coughing recordings are less interesting, but this is a way to build AI that can identify disease symptoms by phone. She said: “It’s a bit disgusting to listen to cough all day!” said Microsoft anthropologist Gray, the job is easily misunderstood. Listening to people’s coughs all day can be disgusting, but it’s also the way doctors spend their days. She said: “We don’t think this is a chore.”

Ms. Hernandez’s job is to help doctors do their jobs, or maybe one day, replace them. She is proud of this. Shortly after complaining about the project, she pointed out that colleagues in the office said: “We are all masters of cough diagnosis.”

“I have had enough”

Uncover the secrets of AI training: In addition to experts and help the AI ​​evolution, there are countless office workers around the world

▲Figure 10: Kristy Milland of Toronto worked at Amazon Mechanical Turk for 14 years. This is a crowdsourced data annotation task company, and now she is trying to improve the working conditions of those who do these jobs. /p>

In 2005, Kristy Milland registered her first job at Amazon Mechanical Turk. She was 26 years old and lived in Toronto with her husband, who managed a local warehouse. Amazon Mechanical Turk is a way to make a little extra money.

The first project is Amazon’s own. There will be three photos of the storefront on Mirand’s laptop, and she will choose the one showing the front door. Amazon is building an online service like Google Street View, which needs help picking the best photos.

She earns $0.03 per click, or about $0.18 per minute. In 2010, Midland’s husband lost her job and Amazon Mechanical Turk became her full-time job. In two years, she worked six or seven days a week, sometimes 17 hours a day. She earns about $50,000 a year. Ms. Milande said: “It was enough at the time, but now it is not.”

The work at the time did not really involve AI. For another project, Mirand will extract information from the mortgage file or retype the name and address from the business card photo, sometimes earning only $1 an hour.

About 2010, Miran began tagging AI projects. She has tagged all kinds of data, such as bloody pictures on Twitter (which helps to build AI, helps remove bloody images from social networks), or may be aerial shots somewhere in the Middle East, presumably It is an AI that is being built by the military and its partners to identify the target of the drone.

Middle said that projects from US tech giants are usually paid more than regular jobs, about $15 an hour. But this job has no health care or paid holidays and can be numb or deeply disturbing. She called it “terrible exploitation,” and Amazon declined to comment.

Since 2012, the 40-year-old Mirand has been staying in an organization called TurkerNation, which aims to improve the working conditions of thousands of people working in this type of work. In April of this year, she resigned after 14 years of work.

Middle is studying law school, her husband’s income is $600 less than the monthly rent they pay, which does not include utilities. Therefore, they are preparing for debt. But she won’t go back and label the data. She said: “This is a dystopian future, I have had enough!”