Data workers who help to shape artificial intelligence large language models used by millions are largely unseen by the public, which leads to misunderstandings about how the AI sector works, a report has said.
Data annotation is done by people who often work on contract, helping to categorise and label information such as specific aspects of photos that ultimately become the backbone of AI tools.
“Many annotators viewed their role as a stepping stone to future professional growth,” read the report compiled by Oxford Economics on behalf of Scale AI, which helps technology firms to train AI.
The report predicts that by 2030, the data annotation market is expected to reach a $19 billion market size in the US alone.
“The data-annotation industry directly supported earning opportunities for nearly 200,000 people and an additional 9,000 full-time jobs,” the analysis says.
It says that although the general public does not see the work of data annotators, they play a critical role in finding the nuance needed to make AI tools accurate.
“By embedding context, describing structure and making the sources that are available richer, data annotators can enhance model performance even where available text has been exhausted.”
The findings go against many concerns about AI, showing that humans are still very much a part of the burgeoning technology sector and that jobs are being created as a result.
Yet the fact that the report was commissioned by Scale AI is likely to invite some criticism.
Scale AI and other companies in the industry have a lot on the line with a possible move from the Department of Labour that would make it easier for firms to use contract workers.
According to media reports, the company has settled lawsuits that alleged it underpaid and misclassified contracted workers in breach of Californian laws.
In March, TechCrunch reported that Scale AI was under investigation by the Department of Labour amid allegations that the company breached the Fair Labour Standards Act, which outlines basic rules for wages and work-safety standards, as well as regulations for contracted workers.
That investigation, according to TechCrunch, was later dropped.
Surge AI, a competitor of Scale AI, is facing a class-action lawsuit alleging that it deliberately misclassified “its data annotator workers as independent contractors, rather than employees, thereby denying them the protections and benefits that they are entitled to as employees”.
The Clarkson Law Firm, which represented the employees making allegations against Surge AI, says that the company demanded “off-the-clock” training, contradicting the definition of independent contractors.
In 2024, Clarkson made similar allegations about Scale AI.
Regardless of the litigation, the AI data-labelling industry shows no sign slowing down, and with its recently commissioned report, Scale AI is trying to improve the standing of the contracted workers.
“The results of Oxford’s study confirm what we see every day,” a recent Scale AI blog post reads. “These contributors are highly skilled people, dedicated to their professions, to their families and to their communities.”
