Story perspectives

Meta Pauses AI Initiative After Employee Data Access Concerns

6/24/2026

Summary

Meta Pauses Initiative

Meta paused its Model Capability Initiative after learning employee computer-use data could be accessed internally, though a spokesperson said no improper access was detected.
Nearly 2,000 Meta workers signed a petition demanding cancellation, prompting the firm to let staff opt out of tracking for 30 minutes.
While spending up to $145 billion on AI this year and conducting layoffs, Meta’s changes were described by former employees as exhausting and depressing.

Summary

Meta Breach

Meta CTO Andrew Bosworth warned staff that the Model Capability Initiative failed the review, causing a SEV 2 breach that exposed data in 45,000 Hive tables.
Workers reported the recording software drained laptop batteries and raised internet usage, fueling unrest as the project was paused after March AI mishaps.
Analysts say the breach shows an industry ‘data wall,’ with firms using employee data for AI training as public web data shrinks.

History

Historical Context

In August 2018, a leaked memo revealed Google’s “Project Dragonfly,” a censored search prototype for China. By September, about 1,400 employees signed a petition and Amnesty International condemned the plan, prompting CEO Sundar Pichai to label the effort “exploratory.” In July 2019, Google’s VP Karan Bhatia testified that the project had been publicly terminated after internal backlash.

1 of 3

Meta Pauses Initiative

Meta paused its Model Capability Initiative after learning employee computer-use data could be accessed internally, though a spokesperson said no improper access was detected.
Nearly 2,000 Meta workers signed a petition demanding cancellation, prompting the firm to let staff opt out of tracking for 30 minutes.
While spending up to $145 billion on AI this year and conducting layoffs, Meta’s changes were described by former employees as exhausting and depressing.

1 / 3