The companies that promote AI fail to mention that it is often not underpinned by code, but by humans viewing and tagging unsavoury content. AI could not exist if cheap labour was not outsourced to the Global South.
The biggest lie about artificial intelligence (AI), is that it is a lie of omission. Its egregious impacts on people and places are not told.
According to the convenient narrative that the Broligarchy is selling us, it’s about the clever machines, the zeros and ones of code, and the super-clever algorithm that will create a new order for all.
They sold us the same dream of our data being in the cloud, a soft, floaty place that we couldn’t reach any more than the clouds in the sky.
In reality, the “cloud” has a greater carbon footprint than airline industry. It is also more physical. This is manifested by water-guzzling and extractive datacentres in environmentally challenging areas.
The people are the focus
Let’s get back to the issue of the people. Data labelling is at the core of AI, without which it would not exist. It’s done by humans, not machines, under grueling conditions. Some are in the US, but the majority are in the Philippines and other parts of the Global South. ChatGPT relies heavily on humans to accurately train its data sets. These people are critical to their business, but they are not “tech workers” nor employees of companies such as Meta, Google, or OpenAI. They are instead outsourced workers who do not enjoy the benefits that the tech sector normally enjoys. They don’t get the on-campus meals, free transport, wellness days, or the high salaries. Many of these people are forced to work from home or in internet cafés because they cannot afford a computer.
The tech giants have been making sure that
These workers sign non-disclosure contracts that prohibit them from speaking about the nature of the work they do and their terms and condition. However, more recently, some of these content moderators have begun to agitate for improvement in their plight. This is evidenced by a letter written by Kenyan moderators addressed to President Biden after his term ends in 2024.
Their letter gives a chilling insight to the state of their lives and careers:
As tech workers, we’re proud to be a part of the development and training for world-class emerging technology – and, most importantly, making it safe. This work is done at great risk to our health, lives and families. The US tech giants are exporting their most dangerous and difficult jobs to other countries. The work is mentally, and emotionally, draining. We clean Facebook, TikTok, and Instagram to make sure these important platforms are not flooded with hate speech or incitement to violent acts. We label images and texts to train generative AI software like ChatGPT for OpenAI. Our work involves watching murders and beheadings as well as child abuse and rape. Pornography and bestiality are also part of our work. Many of us do this for less than $2 an hour.”
Hundreds beheadings
Brilliant documentary called
The cleaners premiered at Sundance Film Festival in 2019 and focuses on content moderators working in the Philippines. It intersperses their personal stories with the geopolitical implications.
I find it tragic to hear a woman calmly talk in the documentary about hundreds of beheadings, as if this was something that any human being could tolerate for $2 an hour. The intercutting between the stories of the content moderators and the cheesy Facebook videos that show Mark Zuckerberg raving about how great Facebook is for the entire world is fascinating.
It’s not as if Zuckerberg is unaware of the appalling working practices and abuses. When asked what he thinks about it, he replied:
In a leaked recording of a company meeting“I think some of the reports are a bit overdramatic”.
Sarah T Roberts, a writer and researcher who spent more than 10 years researching the working conditions and lives of content moderators across the globe for her book.
Behind the screen: content moderation in social media’s shadowsdocuments their lives in great detail. One of her interviewees named “Melinda” described her work as that of a “sin-eater,” which encapsulates the economics of how wealthy tech owners in America exploit the labour of poor people around the world. The sin-eater, who was often a poor community member, would receive financial compensation for taking on another’s sins by eating. In this way, those in economic hardship were more likely than others to ingest other people’s sins. Melinda identified strongly with this forgotten and tragic figure of British legend and not with anyone in the tech or social media sector.
Social and racial justice
Lack of social and racial justice is not limited to the Global South. This work is also done in the shadows by big tech in the US and it’s not a new development. In 2014, a video journalist named Andrew Norman Wilson lost his job at Google after he exposed a shadow team tasked with scanning Google Books. In his
In a YouTube video, he describes a Google hierarchy delineated by different colored badges.
Contractors like himself wore red badges. Interns wear green badges. Full-time Google employees wore white badges, and they were entitled to all of the perks that come with working there.
He discovered that the building next to him was occupied by a class of workers, mainly people of color, Latinos and Blacks. Their shifts started at 4am and lasted to 2.15pm very much like factory workers, the antithesis of the freewheeling work-from-home, come-in-when-you-like tech workers. These workers were not allowed anywhere else on the Google Campus except the office that they worked in called Building 31459.
These workers were not provided with corporate backpacks, mobile phones, thumb drives or any other opportunity for social interaction. Most Google employees were unaware of the yellow-badge classes in building 31459, and many didn’t even know that they existed. Wilson learned to his cost that they were forbidden from discussing their work with any other Googlers. He was fired after he tried to video them leaving the office and for trying speak to some of them.
When Wilson approached one worker, the woman immediately used her mobile phone to call Google Security – as instructed. Wilson’s contract abruptly ended a few hours later and he was escorted out of the building. Wilson, a philosophy and political major, was unable to understand why Google treated people of color as third-class citizens, with their yellow badges and doing critical work that was both unacknowledged, and diminished.
AI Hype
Authors Emily M Bender & Alex Hanna explore these topics in their book, “AI Hype” ]
The AI con: How we can fight Big Tech’s hype to create the future that we wantAccording to the authors “AI hype is used to cover up employers’ actions that lead to the destruction of jobs and workplaces behind shiny claims of techno-optimism.
It spins a future vision where automation will free up people to do interesting tasks while machines will take over the tedious ones. When we look behind the scenes, we find that automation is used as a tool against workers, and is used as a cost-saving measure by employers. This leaves workers with the task of cleaning up the mess, which is devalued, less engaging, and less fulfilling, or worse, is outright painful to do.
While it is convenient to blame Big Tech, they must bear the brunt of the responsibility. To be human, surely, is to care for other human beings?
We cannot pretend when we use these platforms that we are not doing this at a real human cost for people who are least able defend their rights.
This is a fact that we cannot ignore – it has been going on for more than two decades. What are you going to do about your part?
and
Sebastian Klovig Skelton.
And
Sebastian Klovig Skelton.
