Google cuts 200+ elite AI contractors amid unionization efforts

Google has laid off more than 200 contractors who work on improving its AI products, including Gemini and AI Overviews, in at least two rounds of cuts last month. The layoffs come amid an ongoing dispute over pay, working conditions, and alleged retaliation against workers attempting to unionize at outsourcing company GlobalLogic, which is owned by Hitachi.

What you should know: These AI raters are highly skilled contractors responsible for training Google’s chatbots and search features to provide more human-like responses.

• Most raters are required to have master’s degrees or PhDs and include writers, teachers, and creative professionals who evaluate and rewrite AI outputs.
• The work involves rating responses from Gemini, improving AI Overviews (Google’s search summaries), and creating prompts to train the models.
• GlobalLogic employs thousands of these contractors, with most US-based workers handling English-language content.

The big picture: Workers allege the company is systematically replacing them with AI while suppressing their efforts to organize for better treatment.

• Internal documents viewed by WIRED suggest GlobalLogic is using human raters to train an AI system that could automatically rate responses, potentially replacing the human workforce.
• Despite requiring advanced degrees and handling what they describe as “skilled and high-stakes” work, eight workers told WIRED they face underpayment, job insecurity, and poor working conditions.

Key details about the layoffs: The cuts appear targeted at workers who raised concerns about workplace conditions.

• Andrew Lauzon, terminated on August 15, was told the reason was “ramp-down on the project” but received no further explanation after joining in March 2024.
• Two workers have filed complaints with the National Labor Relations Board, alleging unfair termination—one for raising wage transparency issues, another for advocating for coworkers.
• In July, GlobalLogic mandated return-to-office for Austin, Texas workers, directly impacting those with financial constraints, disabilities, or caregiving responsibilities.

The unionization effort: Workers began organizing in late 2023 but faced increasing retaliation from the company.

• A WhatsApp group called “Super Secret Secondary Location” grew to around 80 members discussing workplace issues.
• By December 2023, their Alphabet Workers Union chapter had 18 members, growing to 60 by February after a pay survey was circulated.
• The company banned use of social channels during work hours in February, which workers say was retaliation for discussing pay parity and organizing.

Pay disparities fuel worker frustration: The company created a two-tier system with significant wage gaps for identical work.

• GlobalLogic’s direct “super raters” earn $28-$32 per hour, while third-party contractors doing the same work earn just $18-$22 per hour.
• “Generalist raters” without advanced degrees handle similar AI rating work but report no notable pay increases despite being assigned “more demanding” projects.
• Alex, a generalist rater, estimates roughly 80% of project workers remain on contract without benefits or paid time off.

Working conditions concerns: Employees describe increasingly stressful environments that prioritize speed over quality.

• Task timers were reduced to five minutes, with Alex noting workers are “sacrificing quality at this point” and focusing “more on the timer than anything else.”
• The company has threatened job losses for workers who don’t meet speed metrics, creating what workers describe as an “oppressive atmosphere.”
• Social channels that helped remote workers “feel less robotic and more human” were banned, eliminating important connection points for the distributed workforce.

What they’re saying: Labor researchers say this pattern of retaliation is common in outsourcing companies.

• “This is the playbook,” says Mila Miceli of DAIR Institute, an organization that works with AI data workers globally. “We have seen this in other places, almost every outsourcing company doing data work where workers have tried to collectivize and organize—this has been difficult. They have suffered retaliation.”
• Ricardo Levario, a fired organizer, described building “the movement underground” before being terminated for allegedly violating social spaces policy four days after filing a whistleblower complaint.
• “We can’t really organize—we’re afraid that if we talk we’re going to get fired or laid off,” says Alex about the current atmosphere.

Google’s response: The tech giant distanced itself from responsibility for the contractors’ employment conditions.

• “These individuals are employees of GlobalLogic or their subcontractors, not Alphabet,” said Google spokesperson Courtenay Mencini.
• Google stated it “audits the companies we work with against our Supplier Code of Conduct” but maintains that GlobalLogic is responsible for employment and working conditions.
• GlobalLogic declined to comment on the allegations.

Global context: Similar organizing efforts are emerging among AI workers worldwide.

• Kenyan AI data labelers formed the Data Labelers Association this year to fight for better pay, working conditions, and mental health support.
• Content moderators from Kenya, Turkey, and Colombia formed the Global Trade Union Alliance of Content Moderators in April.
• The pattern suggests growing labor unrest among the largely invisible workforce that trains and maintains AI systems globally.