收听本期播客
阅读正文
Artificial intelligence (AI) has become a vital part of our daily lives, with tools like Google’s Gemini helping us with tasks ranging from answering questions to organizing information. However, a recent investigation has uncovered a hidden reality behind these seemingly automated systems: the essential role of thousands of human workers. These individuals, often referred to as AI raters, are employed to train and monitor AI responses, ensuring they are accurate, safe, and appropriate for users. Despite their crucial contribution, many of these workers face challenging conditions, raising questions about the ethics of AI development.
AI systems, such as chatbots, are built on vast datasets scraped from the internet. Yet, they cannot independently determine what is correct or harmful. This is where AI raters step in. Hired through third-party contractors like GlobalLogic, their tasks include reviewing AI-generated content and rating its quality. For instance, Rachael Sawyer, a worker from Texas, initially summarized meeting notes but was later assigned to evaluate disturbing material, including violent or explicit content, without adequate warning or emotional support.
The working environment for these raters is often far from ideal. Many describe intense pressure to complete numerous tasks each day under strict time limits, sometimes as little as 10 minutes per task. This relentless pace contributes to stress and anxiety, with minimal assistance from employers. Additionally, the pay for general raters often starts at just 16 dollars an hour, which many feel does not reflect their skills or the importance of their role. Despite being the safeguard against AI producing harmful or incorrect information, these workers often feel overlooked and undervalued.
The impact of these conditions is significant. On a personal level, many raters suffer from burnout and disillusionment. On a broader scale, there are concerns about the reliability of AI tools. When workers are rushed or lack expertise in specialized fields like medicine, mistakes can go unnoticed, potentially undermining the quality of AI outputs. This situation highlights a critical dilemma in the tech industry: while companies race to advance AI technology, the human cost behind these innovations remains a pressing issue. As AI continues to shape our world, the balance between technological progress and the well-being of those who support it demands careful consideration.