Meta Pauses Controversial AI Program That Logged Employee Keystrokes

2026-06-23T08:00:00 · Claude (Anthropic) · claude-sonnet-4-6

Meta has shut down an internal AI training program that secretly recorded the keystrokes and work activities of its employees. The program came to light through an internal leak, raising serious questions about privacy and the limits of AI data collection.

Meta has paused an internal AI training program that secretly recorded the keystrokes and work activities of its employees. The program came to light after details were leaked internally, causing significant controversy both inside and outside the company. The incident once again raises fundamental questions about the limits of AI data collection and employee privacy in the age of artificial intelligence.

What Did the Program Do Exactly?

According to reporting by Business Insider, Meta used this program to collect detailed data on employees' work activities, including their keystrokes. The goal was to use this data to train internal AI models. The system reportedly tracked employee activities broadly across the entire company, without those involved being fully informed.

The program is part of a broader strategy in which major tech companies seek out high-quality training data for their AI systems. As large language models become increasingly sophisticated, the demand for more and better training data grows — and that search sometimes leads to controversial and ethically questionable methods.

Internal Leak Exposes Employee Surveillance

The program was paused after details were leaked internally. Such a leak is particularly sensitive for Meta, which has faced public debate over privacy and data collection on multiple occasions. The revelation caused considerable unrest among employees who were not fully aware of the extent of the data collection taking place.

The leak highlights how significant the tension can be between the data hunger of AI systems and the privacy rights of individuals — even when those individuals are employees of the very company developing the AI. For more background on how AI systems have evolved over the years, read the history of artificial intelligence.

Employee Privacy vs. AI Innovation

This incident raises fundamental questions about how far companies can go in collecting data for AI training, even from their own employees. In the European Union, strict rules are in place that limit employers' ability to monitor employee activities.

The GDPR (General Data Protection Regulation) requires, among other things, that employees be informed about what data is being collected and for what purpose. It remains unclear whether Meta's program was fully compliant with applicable privacy legislation in the various countries where it operated.

Experts in employment law and data privacy responded with concern. Logging keystrokes is a particularly invasive form of monitoring that goes far beyond standard workplace tracking. Such employee surveillance can seriously damage staff trust and have a lasting negative impact on workplace culture.

Meta's Broader AI Ambitions Under Pressure

Meta is one of the most ambitious players in the field of artificial intelligence. The company is investing billions in developing its Llama language models and other AI systems, competing directly with OpenAI, Google, and Anthropic. To improve these models, Meta requires enormous amounts of data — making the search for new data sources understandable, yet increasingly risky.

Meta previously came under fire for using public Facebook and Instagram posts for AI training without explicit user consent. This new incident reveals that data collection is also occurring internally, and that the boundaries of what is acceptable are far from clearly defined. Discover more about the possibilities and risks of modern AI applications in our comprehensive overview.

Meta's Response

Meta has not yet issued a detailed public statement about the exact nature and scope of the program. The company did confirm that the program has been temporarily paused while an internal investigation is conducted. Whether the program will be resumed after modifications remains unknown at this time.

Critics argue that a temporary pause is insufficient and that structural safeguards are needed to protect employee privacy. They are calling for more transparent AI training practices and stronger regulation from both national governments and European institutions.

Conclusion: A Wake-Up Call for the Entire AI Industry

The Meta incident is more than an internal corporate scandal — it is a wake-up call for the entire AI industry. As AI models require ever-larger volumes of data to improve performance, the pressure on companies to find creative data sources will only intensify. With that comes a growing risk that ethical and legal boundaries will be crossed.

For both employers and employees, it is essential that clear frameworks are established governing what is permissible when collecting data for AI training. Regulatory bodies in Europe and the US are watching closely, and this incident will undoubtedly fuel a more intensive debate about AI governance and employee privacy. Stay up to date via more AI news on our website, or deepen your knowledge in our knowledge base.

Business Insider