Meta tracks employee usage

- Meta is reportedly tracking employee browsing and usage on sites like Google, LinkedIn, and Wikipedia to train AI models. - The program allegedly collects browsing activity and keystrokes while claiming safeguards prevent evaluation use. - The practice highlights how internal operating changes are being bent into AI training pipelines, raising governance and trust questions (cnbc.com).

Meta is tracking some employees’ clicks, keystrokes and browsing on work computers to collect training data for its artificial intelligence systems. (cnbc.com) CNBC reported on April 22 that Google, LinkedIn and Wikipedia are among hundreds of websites and apps included in the effort, based on internal messages the outlet reviewed. Reuters first reported on April 21 that the internal tool is called Model Capability Initiative, or MCI. (cnbc.com) (reuters.com) Reuters reported that the software is being installed on U.S.-based employees’ computers and can capture mouse movements, clicks and keystrokes. CNET, citing Reuters, reported that the system can also take screenshots of what is on workers’ screens. (reuters.com) (cnet.com) Meta told staff, according to Reuters, that the data will be used to train “AI agents” that can carry out computer-based work tasks. PCMag described the goal as teaching systems to imitate how people move through software and complete routine actions on a computer. (reuters.com) (pcmag.com) CNBC reported that Meta told employees the collected data would not be used for performance reviews or other evaluations. The same report said the program still turns ordinary office activity into raw material for model training. (cnbc.com) The project lands as Meta accelerates spending and restructuring around artificial intelligence. CNBC reported on April 8 that Meta had just introduced a major new model called Muse Spark under chief AI officer Alexandr Wang, and CNBC reported in March that the company was also weighing deep job cuts tied to rising AI costs. (cnbc.com 1) (cnbc.com 2) Meta has also been shifting work previously handled by people toward automated systems. CNBC reported on March 19 that the company was rolling out more advanced artificial intelligence for content enforcement while reducing its use of some third-party vendors and contractors. (cnbc.com) The immediate dispute is not whether companies monitor work devices, which many do, but whether Meta is repurposing that monitoring into a model-building pipeline. Reuters and CNBC both reported that Meta framed the system as an artificial intelligence training tool rather than an employee-surveillance program. (reuters.com) (cnbc.com) Meta’s response, as reported by CNBC, is that safeguards are in place and the data is not meant for employee evaluation. The question now is whether workers, regulators and courts treat that distinction as enough once the tracking is built into everyday work. (cnbc.com)

Get your own daily briefing

Scout delivers personalized news, insights, and conversations tailored to your role and industry.

Download on the App Store

Shared from Scout - Be the smartest in the room.