Meta fires 7,800 employees, uses their work to train AI models
Leaked audio reveals Zuckerberg's strategy to cut costs and replace human roles.
A leaked audio recording from a Meta staff meeting, published by More Perfect Union, reveals Mark Zuckerberg admitting the company actively trained its internal AI models on the work of employees it was already planning to fire. The layoffs—7,800 employees, or 10% of the workforce—coincided with the leak. Meta gave affected staff a one-month notice but kept the names secret until the last minute.
Zuckerberg explained that Meta skipped hiring external contractors to save money, instead using the expertise of its own highly skilled engineers to feed AI training models. He claimed Meta employees have higher average intelligence than contractors, making this approach faster and more effective. This strategy signals a broader industry trend: major tech companies cutting operating costs by using their own workforce's data to train AI systems that will eventually replace human roles.
- Meta laid off 7,800 employees (10% of workforce) after a one-month notice period.
- Leaked audio shows Zuckerberg admitted using employees' daily work to train AI instead of hiring contractors.
- Zuckerberg claimed Meta engineers' higher intelligence made this training approach faster and cheaper.
Why It Matters
Tech giants are now using employee data to train the AI that will replace them—a direct cost-cutting strategy.