Skip to content

Unseen Masters Crafting the Future: The Hidden Shapers of Artificial Intelligence

Data-related work is rapidly becoming one of the most unstable job sectors in Kenya and Germany. Strategies to tackle this issue span from legislative regulation to labor activism, as demonstrated at re:publica 25.

In both Kenya and Germany, the field of data work is increasingly becoming a precarious profession....
In both Kenya and Germany, the field of data work is increasingly becoming a precarious profession. The methods employed range from legislative intervention to labor activism, as evident during the Re:publica 25 event.

Unseen Masters Crafting the Future: The Hidden Shapers of Artificial Intelligence

Seated in a modest apartment in Kenya, data operator Joan Kinyua works from her bed, a testament to both her dwellings' size and her relentless work ethic. With no time for breaks apart from fueling her computer and herself, her workday stretches from midday until the early hours of the morning. This self-funded endeavor, she recalls, is the price of performing her job as a data operator.

For over five years, Kinyua has been instrumental in training artificial intelligence (AI) software. Her expertise lies in teaching self-driving cars to recognize various road signs, as well as guiding the Roomba, the popular robot vacuum cleaner, through its tasks. Kinyua has found herself a part of the gig economy, a vast informal job market where temporary tasks are assigned through online platforms. The concept was initially confined to the Uber drivers, but the field has expanded significantly.

Unfortunately, the gig economy often equates to inhuman working hours, insufficient breaks, insecure contracts, and excessive performance pressure. Kinyua's experience mirrors this reality, as the more she produced, the more she was required, and the less she was compensated, it seemed at the time. During the Christmas season, her company abruptly terminated her contract, leaving her unemployed with her infant child at home.

Kinyua is now the president of the Data Labelers Association, a union for data operators in Kenya and Nairobi. The union's members, whom she refers to as the "invisible architects behind AI," are asserting their labor rights. Kinyua recounts that, initially, they had to educate themselves on these matters, such as the illegality of employment without a contract.

Sitting beside Kinyua on stage was Andreas Hähnisch, deputy chairman of the German TikTok works council. This forum has existed since 2022, and while Hähnisch considers his working conditions "pretty good," he notes that content moderators face "less favorable conditions." This is the very reason why he joined the works council.

Many migrant workers in this sector, particularly freelancers, often don't grasp their labor rights. They may not know, for instance, that poorly performing an employee in Germany makes dismissal difficult or that they are not obligated to sign a mutual termination agreement. Platforms exploit this precarious situation, with technology giants in Germany being no exception.

To foster change, information about workers and their rights could prove valuable. If workers take control of this data, it could provide them with new opportunities and even lead to union level advancements, suggests data anthropologist Alexandra Mateescu. For example, Coworker.org is a program that supports workers in understanding opaque pay models.

As large numbers of robots populate the world, approximately third-highest in Germany, digitization and its impact on working society has become a front-and-center issue. According to Federal Minister of Labor Barbara Bas (SPD), these developments require increased attention on employee data protection to create legal certainty. The Ministry presented a draft bill on employee data protection in the fall of 2024. Bas recently confirmed her intention to revisit this topic at the re:publica conference.

Kinyua conceded that while some initiatives aim for the full automation of data work, human interactions and data workers will still be essential. As such, precarious jobs and opportunities for change will persist. Emphasizing the importance of unions, Hänisch encourages any worker to establish a works council or join one, stating, "It's never too late."

  1. Joan Kinyua, who has been instrumental in training artificial intelligence software, now serves as the president of the Data Labelers Association, a union for data operators in Kenya and Nairobi.
  2. In the gig economy, where temporary tasks are assigned through online platforms, workers often face inhuman working hours, insufficient breaks, insecure contracts, and excessive performance pressure, as Kinyua's experience mirrors.
  3. With the increase in robots worldwide and its impact on working society, particularly in Germany where they are third-highest, there is a growing need for employee data protection to create legal certainty, as suggested by Federal Minister of Labor Barbara Bas (SPD).

Read also:

    Latest