Meta Reassigns 7,000 Employees to AI Task Force Amid Layoffs
Meta has reassigned 7,000 employees to new AI initiatives as part of its strategic shift, coinciding with layoffs affecting 8,000 staff members.
Meta has reassigned 7,000 employees to new artificial intelligence (AI) initiatives, coinciding with layoffs affecting 8,000 staff members. The reassignments are part of a strategic shift to bolster the company's position in the AI sector.
Employees were informed of their new roles via email, with many joining the Applied AI (AAI) group, established earlier this year and led by engineering vice president Maher Saba under the supervision of Chief Technology Officer Andrew Bosworth. Other teams, such as the "Agent Transformation Accelerator" and "Agent Data and Optimization," are also part of this initiative.
The selection process emphasized strong performance and technical abilities. An internal email stated, "This is a reflection of your impact," and noted that recipients were identified as individuals who could make a significant contribution to the team.
While some employees view the reassignments as a positive opportunity, others have expressed uncertainty about their new roles. Discussions on platforms like Blind, an anonymous workplace app, reveal questions about the specific functions of AAI. Some anticipate that the roles may involve data labeling tasks, such as manually tagging images or correcting chatbot responses.
Meta's AI models have lagged behind competitors like OpenAI and Google. To address this, the company is enlisting its employees to enhance AI training. CEO Mark Zuckerberg highlighted this approach in an internal meeting, stating that the average Meta employee possesses "significantly higher" intelligence than external contractors, making them well-suited for this task.
The company has also introduced internal tools like the Model Capability Initiative to track employees' keystrokes and mouse movements, aiming to train AI models on real examples of task completion. An internal announcement noted that AI models still lack the knowledge to outperform humans in technical tasks such as coding.
A Meta spokesperson declined to comment on these developments.