Reuters reports that Meta is dialing back parts of a program that collected employee mouse movements, clicks and keystrokes for use as AI training data, according to an internal memo authored by Stephane Kasriel. The memo, cited by Reuters, says new controls let employees collection for up to 30 minutes and request exemptions. Kasriel also said the team introduced "several optimizations" to reduce battery and data impacts after staff complaints that the software increased home internet usage and drained batteries. Reuters notes the tracking was installed last month on US-based employees' computers as part of a broader initiative to develop AI agents.
What happened
Reuters reports that Meta has scaled back elements of an employee-tracking program that captured mouse movements, clicks and keystrokes for use as AI training data, according to an internal memo authored by Stephane Kasriel, a vice president in Meta's Superintelligence Labs unit. The memo, cited by Reuters, says new controls now allow employees to data collection for up to 30 minutes and to request exemptions. The memo also says the team introduced "several optimizations" to reduce impact on battery life and data usage after employees reported spikes in home internet consumption and battery drain. Reuters reports the tracking software was installed last month on US-based employees' computers as part of a broader effort to build AI agents.
Technical details
Reporting indicates the initiative captured low-level interaction signals such as mouse movement, clicks and keystrokes for model training. Reuters' coverage does not publish technical specifications, telemetry schemas, retention windows, or which internal models would consume the data. The internal memo quoted by Reuters frames the adjustments as controls and optimizations rather than a full suspension.
Industry context
Industry observers have repeatedly highlighted privacy and operational risk when workplace telemetry is repurposed for AI training. Companies that instrument employee endpoints for product research or automation often face tradeoffs between signal fidelity and employee privacy, especially when keystroke-level data is involved. Editorial analysis - industry pattern: firms that encountered staff pushback have typically added opt-out mechanisms, shorter capture windows, and clearer data-use policies to reduce legal and morale friction.
Context and significance
For practitioners, telemetry such as mouse trajectories and fine-grained input logs can be valuable for building agentic interfaces and behavior cloning, but they raise elevated privacy and compliance questions compared with aggregated telemetry. Editorial analysis: teams planning similar data-collection experiments should anticipate employee concerns around personal data on work devices, network costs for remote employees, and battery impact, and budget time for consultative rollout and clearer documentation.
What to watch
Observers will likely track whether Meta publishes a formal privacy notice or updated internal policy detailing retention, anonymization, and model-access controls, and whether regulators, works councils or employee groups escalate complaints. Reuters' report does not include a public statement from Meta beyond the internal memo excerpt.
Scoring Rationale #
The story is notable for practitioners because it highlights a practical privacy and operational tradeoff when using employee endpoint telemetry for AI training. It is not a frontier-model release, but it matters for teams building agentic interfaces and for workplace data governance.
Practice with real Ad Tech data
90 SQL & Python problems · 15 industry datasets
[Active Search Campaigns by BudgetEasy](/problems/sql/active-search-campaigns-by-budget)
[High CPC Clicks & Poor Landing PagesMedium](/problems/sql/high-cpc-clicks-poor-landing-page)
[Campaign ROAS by Attribution ModelHard](/problems/sql/campaign-roas-by-attribution-model)
250 free problems · No credit card