Meta pauses AI training effort tied to employee digital activity, according to market report

Meta Platforms has halted an internal AI training program that involved using employee digital activity, according to a report carried by Yahoo Finance on June 23, 2026.

The market report characterizes the effort as an AI training initiative that relied on observing or using workplace digital behavior by employees. It says Meta responded to questions about the program.

In its reply, a Meta spokesperson indicated the company has paused the program, the report states, and did not characterize the underlying activity as continuing operations in its current form.

The report does not provide granular details on the program’s scope, such as what specific employee systems were involved, how long the training used such data, or whether any outside parties were connected to the training workflow.

It also does not clarify whether model updates derived from the halted program are still in use for public-facing products, or whether Meta has taken additional steps to retrain models, purge data, or audit outputs.

Meta’s statement, as presented in the report, centers on the immediate decision to stop the effort. No further technical disclosure on controls, data retention, or internal governance was included in the market item.

For Meta, the question is larger than a single program, because AI training practices increasingly affect regulatory compliance, privacy expectations, and employee-data boundaries. Model training that touches internal activity can raise internal policy and external expectations, even when the company frames the use as improving systems.

Still, what remains unclear is how far the pause goes. It is not described whether Meta halted all related initiatives, changed the underlying data collection approach, or is conducting an internal review before deciding what, if anything, resumes. Those specifics will likely determine whether this becomes a one-off adjustment or a broader shift in training governance.