Hundreds of Google AI Employees Laid Off Amid Struggles Over Workplace Conditions

Over 200 contractors involved in assessing and enhancing Google’s AI products were unexpectedly laid off in at least two rounds of job cuts last month. This action follows ongoing disputes concerning pay and working conditions, as reported by workers who talked to WIRED.

In recent years, Google has shifted its AI evaluation work—which entails assessing, editing, or rephrasing the Gemini chatbot’s replies to ensure they sound more natural and “intelligent”—to thousands of contractors working for Hitachi-owned GlobalLogic and other outsourcing firms. Most evaluators at GlobalLogic are located in the US and handle English-language materials. Similar to content moderators who manage and categorize content on social platforms, these professionals apply their expertise and judgment to instruct chatbots and other AI tools, including Google’s AI Overviews feature, to formulate appropriate responses across various topics. Workers claim that the latest layoffs coincide with efforts to suppress their protests regarding issues like compensation and job stability.

These employees, often recruited for their specialized expertise, are typically required to hold a master’s degree or PhD to enter the super rater program, encompassing a diverse background of writers, educators, and individuals from creative sectors.

“I was just let go,” says Andrew Lauzon, who received an email about his termination on August 15. “I inquired about the reason, and they mentioned ramp-down on the project—whatever that implies.” He joined GlobalLogic in March 2024, where he was responsible for rating AI outputs and creating various prompts for the model.

Lauzon emphasizes that this action highlights the precarious nature of content moderation positions. He asserts that GlobalLogic has been increasingly laying off employees this year. “How can we feel secure in our jobs knowing we might be dismissed at any time?” he remarked.

Remaining workers at the company express growing worries that they may be training their own replacements. Internal documents reviewed by WIRED indicate that GlobalLogic appears to be using these human raters to train a Google AI system capable of autonomously rating responses, aiming to substitute them with AI.

Concurrently, the company is exploring methods to dismiss current employees while continuing to recruit new ones. In July, GlobalLogic enforced a return-to-office mandate for its employees in Austin, Texas, as per a notice obtained by WIRED. This has adversely affected several workers who cannot afford the commute due to financial issues or are unable to work because of disabilities or caregiving responsibilities.

Despite engaging in what they consider skilled, high-stakes work, eight workers who spoke with WIRED report being undercompensated and facing job insecurity along with poor working conditions. These alleged circumstances have diminished worker morale and hindered their efficiency, according to sources. Some contractors attempted to unionize earlier this year but say those efforts were stifled. They now allege that the company has retaliated against them. Two employees have filed a complaint with the National Labor Relations Board, claiming they were unlawfully terminated—one for raising wage transparency concerns, and the other for advocating on behalf of himself and his colleagues.