Thousands of “overworked, low-wage” people will train Google's AI to look smart, the report reveals
—————
An interesting report from the Guardian reveals the hidden “shadow workforce” of human contractors that are essential for developing and mitigating Google's AI models, including Gemini Chatbot. Often said to be overworked and supercharged, these workers are responsible for a variety of roles that are important to the functioning and safety of AI.
This article highlights the difficult conditions faced by these evaluators who are contracted through companies like Globallogic. Although many were hired for seemingly benign roles, they found that their work involved content moderation and exposure to highly painful or explicit material without prior warning or mental health support.
The report also details the growing disillusionment among these workers who have expressed concern over tight deadlines and loosening safety guidelines. For example, this article focuses on policy changes that allow AI to repeat hate speech, racial slander, and explicit material when it is part of user input. As a result, many of these important workers have grown to distrust the very AI products that they help them build, viewing technology as “technical magic” and products built on the back of the almost invisible human labor force.
