What happened when Meta tried to collect employee data

Meta faced significant backlash when it announced it would begin collecting employee data to train AI models. Employees at the company pushed back against the plan to capture keystrokes and mouse movements, circulating an online petition and urging colleagues to sign. Within two months, Meta had halted the program due to employee concerns and a data privacy breach.

The idea to collect employee data reportedly came from researchers in Meta's AI division, who were largely isolated from the rest of the company. The program, called the Model Capability Initiative (MCI), was announced as mandatory, drawing immediate outrage from employees. Cybersecurity concerns were raised, with a cybersecurity engineer claiming the program seemed "an afterthought" and not secure.

Meta CTO Andrew Bosworth defended the program, stating it was "tightly controlled data" and that it couldn't end up in the "open." Many employees worried the program would extract sensitive data like passwords and credit card information, and some feared it would lead to job losses and replace employees with AI. A petition with over 1,800 signatures was created, and workers used physical flyers to gather more support.

In June, Meta allowed employees to opt out for 30 minutes but ultimately suspended the program after a Meta engineer discovered that the gathered data was accessible to the entire company. Employees reportedly expressed concerns that the company was "the richest yet most incompetent" tech company. The backlash indicates the risks far outweighed the benefits of implementing such a controversial program, raising questions about the ethics of collecting employee data, even if it is technically legal.