2 hours ago
3
Meta employees are fuming over a new internal tool that tracks their clicks and keystrokes — the latest flashpoint in the company’s all-in push on artificial intelligence.
Staffers reacted with alarm after learning the software would monitor mouse movements, typing and on-screen activity to help train Meta’s AI systems, according to internal communications viewed by Business Insider.
One employee wrote, “This makes me super uncomfortable. How do we opt out?” — a comment that quickly became the top response, as others flooded the thread with angry reactions, including a wave of “angry face” emojis.
Employees were reacting to a new internal program first reported by Reuters, which said Meta would train its AI systems by analyzing keystrokes and mouse movements on workers’ computers.
Executives told workers there would be no option to opt out on company-issued devices, while Meta said the program includes safeguards and is limited to work-related applications.
The backlash comes as Meta ramps up a sweeping — and costly — overhaul of its AI strategy, pouring billions into infrastructure, talent and new products in a bid to catch up with rivals like OpenAI and Google.
Meta is also reported to be on the verge of slashing its workforce, with the company planning to cut about 10% of its global staff — or roughly 8,000 workers — starting May 20 and potentially more later this year as it pivots toward AI-driven roles.
The layoffs are part of a broader effort to reshape the company around artificial intelligence, with executives pushing to automate tasks previously handled by human workers, according to Reuters.
“There are safeguards in place to protect sensitive content, and the data is not used for any other purpose,” a Meta spokesperson told Business Insider.
The tracking tool — known internally as the Model Capability Initiative — captures employees’ mouse movements, clicks and keystrokes while they work, and can also take periodic snapshots of what’s on their screens, according to internal memos cited by Reuters.
The software runs only on work-related applications and websites, with the goal of collecting real-world data on how people actually use computers.
Meta says the data will be used to train AI systems to better mimic everyday human behavior — such as navigating menus, using keyboard shortcuts and completing routine digital tasks — areas where current models still struggle.
The company has said the information won’t be used for employee performance reviews and that safeguards are in place to protect sensitive content, though the initiative is part of a broader push to build AI agents capable of handling work tasks autonomously.
Earlier this month, the Financial Times reported that Meta staffers will have the option of chatting with a virtual AI-powered clone of CEO Mark Zuckerberg.
The Post has sought comment from Meta.
English (US)