Meta is suspending its internal AI training program after sensitive data became accessible company-wide, according to screenshots obtained by Business Insider.
Screenshots show that the breach exposed employees’ private conversations, performance data, and transcripts. This incident is classified as SEV 2 on a scale of 1 to 5, with 1 being the most severe.
A Meta spokesperson confirmed the incident and said the company is investigating.
“We have carefully designed this program with privacy safeguards in place, and at this time there is no indication that data has been compromised by Meta employees, but we have suspended the program while we investigate,” the spokesperson said.
In April, Meta announced an AI training program called the Model Capability Initiative (MCI). The aim was to use staff keystrokes and mouse movements as training data to improve the company’s AI models. The program, which is mandatory for most staff, sparked a backlash from employees who were uncomfortable with their data being recorded, Business Insider previously reported.
The breach has sparked frustration within Meta, with employees criticizing the data for not being locked down in the first place, according to screenshots seen by Business Insider.
One employee wrote on Monday that she was “outraged” about the recent breach in a company group, according to screenshots obtained by Business Insider.
“While we see no evidence of malicious access, the fact that this data was not locked down as originally promised is extremely upsetting,” the employee continued.
The breach is the latest in a series of recent security incidents at Meta. Last month, a flaw in the company’s AI chatbot allowed people to take over multiple Instagram accounts. The Information reported that a rogue AI agent caused a serious incident in March.