Meta pauses training of AI that monitored employees' keyboards due to data leak
Meta is suspending its internal AI training program after confidential data was made publicly available to all company employees.
Meta is suspending its internal AI training program after confidential data was made publicly available to all company employees.
Meta is suspending its internal AI training program after confidential data was made publicly available to all company employees.
This is evidenced by screenshots obtained by Business Insider .
One screenshot showed that the leak exposed private conversations, performance data, and text transcripts. The incident was given a severity level of SEV 2 on a scale of 0 to 5, with 0 being the highest threat level.
A Meta spokesperson confirmed the incident and said the company is currently investigating.
“We carefully designed this program with privacy protections in place. And while we currently have no evidence that Meta employees have improperly accessed any data, we are suspending it while we investigate,” a company spokesperson said.
In April, Meta announced the launch of an AI training program called the Model Capability Initiative (MCI) . It aimed to improve the company’s AI models by using its employees’ keystrokes and mouse movements as training data. The program, which is mandatory for most employees, has sparked outrage among employees who were uncomfortable with the recording of their data .
According to screenshots, this leak has caused a wave of outrage within Meta: employees are criticizing the company for not securely restricting access to the data from the very beginning.
“I’m just angry,” one employee wrote in an internal group.
“I don’t see any evidence of malicious access, but the fact that this data wasn’t protected as promised from the start is incredibly annoying,” another employee added.


