Meta has paused an internal program that monitored employees’ computer use and fed the data into artificial intelligence systems, after a staff petition raised concerns about privacy and data handling, according to reporting by the Guardian.
The tool recorded workers’ keystrokes, mouse clicks and on‑screen content, and was intended to supply training data for Meta’s AI models. Around 1,600 employees signed a petition urging the company to stop using the system, the Guardian reported, leading Meta to put the program on hold.
What Meta’s tracking tool did
According to the Guardian’s account, Meta deployed a software tool on staff computers that:
- Logged keystrokes
- Recorded mouse clicks
- Captured the content displayed on workers’ screens
The reporting indicates that this information was then used as training material for Meta’s artificial intelligence systems. Training data in this context means real‑world examples that machine‑learning models analyze to learn patterns, improve predictions or generate text and images.
The Guardian’s coverage did not specify how long the tool had been in place, which teams were covered, or whether any categories of data were excluded from collection. It did, however, state that the system operated broadly on staff devices and that it was linked directly to AI development efforts inside the company.
Employee petition and privacy concerns
The petition, described in the Guardian’s reporting, was signed by about 1,600 workers. Signatories objected to the scale and nature of the monitoring, focusing on two main issues:
Intrusiveness of monitoring – Because the tool captured keystrokes, mouse activity and full screen content, employees were concerned that highly sensitive information could be recorded. That could include personal messages, passwords, or confidential work material that staff did not expect to be used beyond its immediate purpose.
Use of data for AI training – The Guardian reported that the collected data was used to train Meta’s AI systems. Employees questioned whether they had meaningfully consented to this use and how securely the information would be stored, processed and potentially reused.
The petition framed these issues as data privacy risks for employees, rather than as a standard productivity‑monitoring program. The Guardian did not reproduce the full text of the petition, but its summary emphasized concerns about how deeply the tool reached into workers’ day‑to‑day computer activity and how that information might be repurposed.
Meta’s response: program paused
In response to the petition and internal pushback, Meta has paused the program, the Guardian reported. “Paused” in this context means the tool is no longer actively operating to collect new data for AI training, based on the outlet’s account.
The Guardian’s reporting did not include a detailed public statement from Meta outlining:
- Whether previously collected data will continue to be used
- Whether the company plans to modify and relaunch the program
- What changes, if any, will be made to address privacy concerns
The available information is therefore limited to the fact of the pause and the link to employee objections. The Guardian did not report on any formal policy changes or updated internal guidelines associated with the decision.
Why the pause matters inside Meta
Within Meta, the pause affects two overlapping areas described in the Guardian’s report:
Employee monitoring practices – The company has stepped back, at least temporarily, from a form of detailed digital surveillance that included keystroke logging and screen capture. The petition suggests that a significant group of staff saw this as overstepping acceptable workplace monitoring.
AI training data sources – The program was designed to feed employee activity into Meta’s AI training pipelines. Halting it removes one source of real‑world data that engineers and researchers had been using or planned to use.
The Guardian’s reporting did not quantify how important this internal data was to Meta’s broader AI efforts or whether it was experimental. However, the direct link between employee activity and AI training was central to the concerns raised in the petition.
Limited public information and what to watch
Independent corroboration of the details remains limited at this stage. The Guardian is the primary outlet describing the tool, the petition and Meta’s decision to pause the program. Other major outlets had not, as of the time reflected in the available evidence, published detailed accounts of the same system.
Given that constraint, several points remain unclear based on the Guardian’s reporting alone:
- How widely the tool was deployed across Meta’s global workforce
- What categories of data, if any, were filtered out or anonymized
- Whether regulators or external privacy authorities have been notified or involved
Further developments to watch include:
- Any formal public statement from Meta describing the scope of the program and the reasons for pausing it
- Clarification on whether data already collected will be retained, deleted or restricted
- Additional reporting that either confirms or adds detail to the Guardian’s account
Why this development is significant
The Guardian’s reporting shows that Meta has halted a program that blended intensive employee monitoring with AI training, after about 1,600 workers signed a petition objecting to the practice. The pause highlights how internal staff reactions can shape how large technology companies collect and use data, even when those efforts are tied to high‑priority AI projects.
For readers, the key takeaway from the available evidence is straightforward: Meta tested or deployed a tool that recorded detailed employee computer activity to train AI systems; a substantial group of workers objected on privacy grounds; and the company has now paused the program while those concerns are in focus.
