More than 200 contract workers helping Google improve its artificial intelligence systems have lost their jobs in recent weeks, according to a report by Wired. The cuts were carried out by outsourcing firm GlobalLogic, which handles much of Google’s AI rating work. For many of these workers, the layoffs have added to concerns that their efforts were being used to train AI models that could one day make their own roles redundant.
Google fires 200 AI contractors who feared they were training bots to take their jobs
The contractors worked on tasks such as reviewing and editing responses from Google’s Gemini chatbot and polishing AI-generated summaries for Google Search. This work, often described as “AI rating,” required them to check whether the system’s answers were accurate, natural-sounding, and grounded in reliable sources. Many of the so-called “super raters” had advanced degrees and came from professional backgrounds such as teaching, writing, and research.
Despite the specialised nature of their jobs, workers say they faced low pay, tight deadlines, and very little job security. Some allege that GlobalLogic has been using their expertise to train systems designed to automate the very work they were doing. Internal documents reviewed by Wired reportedly show that the company was developing AI tools that could automatically rate chatbot responses. “I was just cut off,” said Andrew Lauzon, who was told by email on August 15 that his contract was ending. “I asked for a reason, and they said ramp-down on the project, whatever that means.”
The cuts also come against a backdrop of worker unrest. Over the past year, some contractors have attempted to organise and push for better pay and transparency. Others voiced concerns about workload pressure, including strict time limits for tasks that left them focusing on speed rather than quality. A group even explored forming a union under the Alphabet Workers Union, but workers claim those efforts were discouraged. At least two have now filed complaints with the US labour board, alleging they were dismissed for speaking up about conditions.
There are also reports of growing disparities within the workforce. Workers directly hired by GlobalLogic earned about $28 to $32 an hour, while contractors brought in through third-party agencies were paid $18 to $22 for the same assignments. Generalist raters, who did not need advanced degrees, were paid even less despite sometimes being pulled into more complex projects.
Some remaining workers say they feel trapped. A few who joined GlobalLogic as full-time employees received better stability, but the majority continue to be on short-term contracts without benefits or paid leave. Many say they now hesitate to raise concerns, fearing they could be next in line for dismissal. “It’s just been an oppressive atmosphere,” one rater said.
Google, meanwhile, has tried to distance itself from the dispute. “These individuals are employees of GlobalLogic or their subcontractors, not Alphabet,” spokesperson Courtenay Mencini said. “As the employers, GlobalLogic and their subcontractors are responsible for the employment and working conditions of their employees.” GlobalLogic has declined to comment.
Labour researchers say the situation shows a wider pattern in the AI industry, where contract workers power the training of advanced systems but often lack recognition, fair wages, and protections. Similar battles are being fought elsewhere, with data labelers in Kenya and content moderators in countries such as Turkey and Colombia also organising for better treatment.
