Artificial Intelligence (AI) companies are increasingly relying on online labor platforms, also known as cloudwork, for tasks like data labeling and content moderation. While this offers AI companies a cost-effective way to train their systems, it raises significant concerns about working conditions and labor rights for the millions of workers registered on these platforms.
AI companies' reliance on online labor platforms
Big names in the tech world - from Amazon to OpenAI - are regularly turning to online labor platforms for their needs. These platforms, including Scale AI, Freelancer.com, Amazon Mechanical Turk, Fiverr, and Upwork, have become instrumental in training AI systems and outsourcing various tasks. This is not exclusive to tech companies; universities, businesses, and NGOs also hire from these platforms. The roles they need filled range from data labeling to content moderation and even IT expertise. These platforms have become the go-to for everything from training new AI systems to ensuring existing ones perform as expected.
These labor platforms are about more than just AI companies' needs. They've become a fundamental source of income for a growing portion of the population. A study by the University of Oxford estimated that over 163 million people have registered on these platforms. Freelancer.com alone boasts a workforce equivalent to the population of the United Kingdom - over 67 million workers. Many of these workers are based in low- to middle-income countries and work for clients worldwide, effectively creating a global workforce.
For AI companies, the appeal of these platforms is obvious. They provide a cheap, often skilled, workforce accessible around the clock. And with fierce competition for jobs, hiring companies usually have the upper hand in negotiating conditions and pay. However, this isn't a rosy situation for everyone. For the workers on the other side, the realities can be harsh. They may spend large chunks of their day searching and applying for jobs, often without pay. Moreover, there's no guarantee that the offered tasks will meet their minimum wage expectations. If any issues arise with a client, there's not always a clear appeal process, leaving them at risk of not being paid at all.
Low scores for cloudwork platforms on fair work principles
A project by the University of Oxford called Fairwork evaluated 15 of these cloudwork platforms. They scored companies based on the principles of fair work - pay, conditions, contracts, management, and representation. None of the platforms evaluated met more than half of these basic standards of decent work. The situation is particularly grim on popular platforms like Fiverr and Freelancer.com, which scored 2 and 1 points, respectively. The average earnings of workers surveyed amounted to a dismal $2.15 USD per hour of work, pointing to the vast gap between the needs of workers and the realities of the platform economy.
The need for international labor standards in the digital economy
Given the international nature of the platform economy, it's essential that the global community comes together to address these challenges. The International Labour Organization is currently discussing the development of standards to ensure workers' rights in the digital economy. While this is a promising step, it won't be effective if national governments don't get involved. So far, discussions around regulating the platform economy have primarily focused on sectors like food delivery and ride-hailing. Yet, it's crucial that policymakers and regulators also consider the unique challenges faced by online remote workers, and put the same energy into finding solutions for this hidden, but very real workforce.
