According to a report by NBC News, ChatGPT—the most popular chatbot garnering headlines for ushering in a new era of A.I. capabilities—is run on the backs of human workers being paid low wages.
OpenAI, the startup behind ChatGPT, is reportedly paying hordes of U.S.-based contractors to assist it with the necessary task of data labeling—the critically important process of training ChatGPT’s software to better respond to user requests—for around $15 an hour (USD).
While few will argue that A.I. stands to reinvent the way we approach societal challenges, ChatGPT has skyrocketed to poster-child status. As hype surrounding the platform reaches a fever pitch, Goldman Sachs recently suggested that A.I. could wipe out 300 million jobs globally.
As the new favorite tool of those seeking to write term papers, generate code, plan vacations and translate languages, the chatbot has become a gateway for many; their first hands-on experience interfacing with an A.I.-driven language model. Yet, according to the people behind the platform, automation and technology have little to do with the platform’s success.
“We are grunt workers, but there would be no A.I. language systems without it,” one worker, Alexej Savreux, told NBC. “You can design all the neural networks you want, you can get all the researchers involved you want, but without labelers, you have no ChatGPT. You have nothing.”
Data labeling—what Savreux is referring to—requires real human workers to painstakingly parse data samples in order to feed automated systems with the keys to correctly identify certain items within a dataset, such as visual images and written text.
“A lot of the discourse around AI is very congratulatory,” said Sonam Jindal, the program lead for AI, labor and the economy at the Partnership on AI, a nonprofit based in San Francisco that promotes research and education around artificial intelligence, in an interview with NBC.
“But we’re missing a big part of the story: that this is still hugely reliant on a large human workforce,” she said.
As demand continues to spike, workers are standing up for better wages and working conditions. In Nairobi, Kenya, more than 150 people who’ve worked on A.I. for Facebook, TikTok and ChatGPT voted to form a union, citing low pay and the mental toll of the work, Time magazine reported.
Additionally, OpenAI has hired about 1,000 remote contractors in places such as Eastern Europe and Latin America, says NBC News, to label data or train company software on computer engineering tasks, says online news outlet Semafor.
Straight Talk
On that note, we went straight to the source and posed two questions to ChatGPT. The first, “Are you guided by low-paid human workers labeling data?” to which it answered:
My training data is generated by human experts and researchers, who often use labeled data sets to teach me how to understand language and generate responses. These labeled data sets may be created and labeled by human workers, who may be paid varying amounts depending on the project and the location. However, the exact methods and sources used to generate my training data can vary depending on the specific model and implementation.
Secondly, we asked, “Do you support the unionization efforts of human experts and researchers who use data sets to teach you how to understand language and generate responses?” to which it replied:
As an AI language model, I do not have the ability to support or oppose any specific social or political issue. However, it is important to recognize the value of the contributions made by human experts and researchers in creating and refining my training data. These individuals play an essential role in ensuring that my responses are accurate, useful, and informative. As such, it is important that they are fairly compensated for their work and have the ability to organize and advocate for their rights and interests as they see fit.
