Meta Trained AI Using Laid-Off Employees' Work

According to internal meeting transcripts released by More Perfect Union, Mark Zuckerberg admitted that he used the work of soon-to-be-laid-off employees to train Meta's internal AI models. Layoffs affecting 7,800 people began this week, as part of the 10% overall workforce reduction announced in April.

Zuckerberg said Meta decided to leverage the expertise of its own employees instead of hiring external contractors. He explained that because Meta employees have higher average intelligence than typical contractors, directly observing the company's engineers' day-to-day work to train code-writing capabilities is far faster and more effective than other approaches used in the industry.

This case illustrates a strategy in which major tech companies train next-generation AI systems using the data and skills of their own workforce, suggesting they are aggressively pursuing both operational cost reduction and replacing human roles with AI.