Reporting by XinZhiyuan
Edit: Time Rayan
AI is repeating the world colonial model, South Africa, Venezuela and other third-world citizens are forced to engage in low-end labor in order to survive, from screws in factories to labeling data, the same destination!
Historically, these countries and regions have been impoverished by the former colonial empire, and European colonialism has been characterized by violent seizure of land, exploitation of resources and exploitation of people.
However, today's AI industry is repeating this model, at the expense of the poor, to make the rich and powerful richer.
Now, these colonized people are reaffirming their culture, their voices and the way they decide the future.
South Africa: Surveillance technology fuels apartheid
How can European and American countries with advanced technology implement digital "colonization" in technologically backward areas like Africa?
In South Africa, even 5 years ago, this kind of "colonization" was not possible.
However, as fiber coverage within South Africa expands and AI capabilities improve, some overseas companies see an opportunity and begin dumping the latest surveillance technology into South Africa.
Under the pressure of a high-crime environment, South Africa has embraced options.
Instead of displaying dozens of video streams at the same time, Proof 360 uses artificial intelligence and other analytics techniques to display only footage of video that triggers security alerts, including systems for license plate recognition and detecting "anomalous" activity.
South Africa has more than 1,100 police stations and more than 180,000 staff, but there are 11,372 registered security companies and 564,540 active security personnel, more than the police and the military combined.
South Africa is not only a high-growth market for surveillance systems, but also a place for improved technology.
"When AI is developed in Europe and the United States and all these places, South Africa is usually the place to test them." Kyle Dicks, a sales engineer at Axis Communications in Johannesburg, said.
Separately, South Africa is building the ABIS National Biometric Database, which will include the faces of every resident and foreign tourist.
Combined with camera upgrades and expanded use of facial recognition in surveillance networks across the country, ABIS could one day allow the government to track the actions of everyone in the country.
A 2015 government report showed that half of South Africa's population lives in poverty, 93 percent of whom are black, and that South Africa's private surveillance machines are fueling digital apartheid.
Venezuela: Low-end digital labor
Is artificial intelligence creating a new colonial world order?
As a third world country, Venezuela's economy has been in trouble for many years. Inflation is soaring and jobs are not stable.
In search of better opportunities, millions of migrants and refugees inside Venezuela have left their countries.
And for most other Venezuelans, leaving the country is impossible.
The root of it all lies in Appen, an AI data annotation platform.
An increasingly unstable wave of crime has trapped them at home, they have lost other jobs, they have to resort to data annotation work, and many families have begun to work full-time on the platform.
Sometimes, parents and children take turns using shared computers, and women take care of housework so that men can work around the clock.
But as Oscarina Fuentes Anaya quickly discovered, the window of opportunity is getting smaller and smaller.
Shortly after the Spare5 shutdown and pandemic outbreak, more and more workers joined the platform, and the number of tasks on Appen began to decrease.
What used to be a reliable 24-hour-a-day task queue, she said, is now getting emptyer and more empty, with work arriving irregularly and randomly.
Appen divides its account into four levels, and users need to complete level 0 and level 1 tasks according to a consistent standard before they can access other work at level 2 and level 3.
Over time, lower-level tasks are almost non-existent, which means that the money received by the creator of the new account is negligible.
The only way to get a job is to buy an existing premium account in the underground marketplace, but people who do so run the risk of being shut down because they violate company policies.
Christina Golden, a spokeswoman for Appen, said it had abandoned the level-based model, but that its projects were still "qualified and therefore not open to everyone."
"We pride ourselves on paying above the minimum wage and adhering to an ethical code." She added.
"During the crisis, hopefully our platform will be a beacon for Venezuelans and provide jobs for those who need it." She said.
This makes Remotasks the next best choice.
While Hive Micro is the easiest service to join, it offers the most disturbing jobs — such as tagging images of terrorists — with the poorest pay.
But as soon as Remotasks Plus was launched, problems began to occur with the system.
Many users quickly realize that their working hours are undervalued, which reduces their weekly earnings, and they are also being asked to adhere to higher standards, with a greater risk of being suspended because they are not fast enough or precise enough.
Simala Leonard, an artificial intelligence student at the University of Nairobi, who worked on the Remotasks project for several months. He said the salaries of the data commentators were "totally unfair".
Simala argues that those who work on "the most basic parts of machine learning" are paid meager, compared to algorithm developers at Google and Tesla's self-driving cars who earn six-figure salaries.
Hossam Ashraf Esmael, a former North Africa community manager, said: "Remotasks treated us as inhuman as if we didn't deserve to make enough money."
Even in this case, Venezuelan worker Ricardo Huggines joined Thermotasks to support his wife and children.
Migrant Workers: "Don't Forget Us"
Appen sends different jobs to different workers, based on the distribution of a range of signals including their location, speed, and proficiency.
Although the people in the group did not know the exact mechanism, they knew that they had received different tasks.
Fuentes worries that one day Appen will abandon her, and she is grateful to Appen.
"I survived because of this platform", other platforms have stopped paying, but Appen has always been there.
"I hope that in 4-5 years, Appen will become a more traditional employer," she says. They know we exist, they know we're sick, they know we need safety and health care."
She has been serving the platform and its clients as an anonymous employee for years, and she wants people to see her face and know her name.
When the sun began to set, Fuentes asked her uncle to take a picture. She held her dog with a smile on her face.
A few weeks later, she sent the photo with a message: "Don't forget us."
Resources:
https://pulitzercenter.org/stories/south-africas-private-surveillance-machine-fueling-digital-apartheid
https://www.technologyreview.com/2022/04/20/1050392/ai-industry-appen-scale-data-labels/