Meta is deploying software on US employees' work computers to monitor keystrokes, clicks, mouse movements, and screenshots in work apps for AI training data. Internal memos reveal no opt-out option, sparking employee discomfort, as the company invests billions in AI while cutting thousands of jobs.
Meta's Model Capability Initiative, outlined in internal memos from Superintelligence Labs, will track US-based employees' and contractors' actions on work-related apps and websites like Gmail, GChat, and Metamate, capturing dropdowns, keyboard shortcuts, and periodic screenshots for context. Phones are excluded. 'This is where all Meta employees can help our models get better simply by doing their daily work,' one memo states, as reported by Reuters and Business Insider. Meta CTO Andrew Bosworth described the vision of AI agents performing tasks under human direction.
Employees expressed strong unease on internal forums. One wrote, 'This makes me super uncomfortable. How do we opt out?' Bosworth replied, 'There is no way to opt out on your work laptop,' prompting shocked reactions including crying and angry emojis. The program targets US staff to avoid EU laws on employee monitoring.
Spokespersons Andy Stone (to Reuters) and others (to CNET) emphasized the data provides 'real examples' of computer use to improve AI agents' handling of tasks like navigation and clicks, with safeguards against sensitive content and no use in performance reviews. This addresses data challenges amid competitors like OpenAI and Google developing similar agents.
Meta plans AI investments over $135 billion this year, alongside layoffs: 8,000 jobs (10% of 79,000 workforce) starting May 20, and 25,000 cuts since 2022. Eric Null of the Center for Democracy & Technology called it highly invasive, urging privacy protections. University of Washington professor Bill Howe noted the data's value despite poor optics.
