Recent headlines have brought to light concerns regarding the sustainability of artificial intelligence (AI) companies like OpenAI and Anthropic in their ability to maintain the rapid pace of data collection required to build more complex AI models. The Wall Street Journal highlighted that these companies have exhausted the entire internet as a data source and are in need of new avenues for data collection. This raises important questions about how data collection and other AI labor practices are being conducted. There is a lack of transparency surrounding data collection and AI system creation, leading to an invisible form of forced labor within the industry.
A new type of laborer known as “Zombie Trainers” has emerged in the AI field. These individuals, including data labelers, content moderators, and image data collectors, unwittingly perform tasks that contribute to the development of AI systems. Captive audiences such as refugees, children, prisoners, and low-wage workers are all considered Zombie Trainers, as they are unaware of the tasks they are performing and the impact on the communities involved. While corporate leaders may argue that these individuals are not unpaid laborers, citing reasons such as being children or prisoners, the lack of compensation for their contributions raises ethical concerns.
Amazon delivery drivers, for example, are highly profitable Zombie Trainers. Cameras installed in Amazon vehicles provide essential data for creating high-definition 3D maps and advancing autonomous vehicle technology. Despite not being explicitly identified as AI trainers in their job descriptions, drivers are unknowingly contributing to the development of AI systems. The partnership between Netradyne and companies like Hyundai enables the use of data collected by commercial truck drivers and delivery drivers without compensating them for their work, highlighting the exploitative nature of this practice.
Children are also unwittingly serving as Zombie Trainers through their interactions with technology. Tech companies collect vast amounts of data from children through various online activities, using it to build AI systems and share or sell it for profit. Parents often lack awareness of the extent of data collection and its consequences, further complicating the issue. There is a need for updated child labor laws to specifically address the use of children as Zombie Trainers and protect their rights. Regulatory agencies must educate themselves on these practices and enforce laws to prevent forced labor in the AI industry.
Governments and regulatory agencies play a crucial role in addressing the issue of forced labor within the AI industry. Child labor laws need to be updated to prohibit the use of children as Zombie Trainers, and guidelines from U.S. Homeland Security should be implemented to combat forced labor through fraud. It is essential for society to collectively decide where they stand on forced labor practices and advocate for ethical AI labor practices. The narrative around AI labor needs to shift towards transparency and respect for the rights of individuals contributing to the development of AI systems.
