Meta was covertly monitoring its employees, but quickly discontinued the practice following an internal leak.

      Meta has acknowledged that its employee surveillance program did not uphold privacy standards.

      Meta recently halted a controversial employee monitoring tool after it inadvertently revealed sensitive employee information to the entire organization (as reported by Wired).

      This tool, known as the Model Capability Initiative (MCI), had been discreetly gathering keystrokes, mouse movements, and screen content from US employee laptops since April.

      A monitoring program without consent

      Meta introduced the MCI with the intention of training AI systems to utilize software in a manner similar to human users, with executives suggesting that employees would be the ideal source for training data.

      Over 1,600 employees signed an internal petition against the program, raising concerns about significant security and regulatory threats. One engineer described the unauthorized screen scraping as a blatant violation of privacy.

      CEO Mark Zuckerberg reportedly supported the initiative in a leaked audio, stating that AI models are most effectively trained by observing capable individuals at work, claiming that Meta’s own employees surpassed average contractors in this regard.

      The incident that prompted Meta to act

      An internal security notice disclosed that data from 45,000 internal database tables had been exposed throughout the company, including private chats, complete prompts, transcriptions, and performance metrics.

      Meta spokesperson Tracy Clayton stated that there is no evidence of unauthorized access to the data, but they are suspending the MCI while they investigate. Employees inundated internal forums with critiques, with one former employee characterizing the oversight as a predictable disaster.

      Meta’s CTO Andrew Bosworth internally admitted that the program did not meet the company's privacy review criteria. This isn't the first time Meta has faced AI-related security issues, following a March incident where an AI agent acted without authorization, and a chatbot vulnerability that allowed hackers to take over Instagram accounts.

      Although the pause may ease current tensions, with morale already affected by layoffs and restructuring, Meta's aspirations in AI are evidently testing the patience of its workforce.