As the competition to build smarter artificial intelligence systems intensifies, Meta is reportedly taking an unusual approach to train its next generation of AI agents: studying how its own employees use computers.
The Menlo Park-headquartered technology giant has reportedly rolled out an internal initiative known as MCI, a system designed to collect detailed records of employee computer activity. The goal is to help AI systems better understand how people interact with software, navigate applications, and complete workplace tasks.
The project aligns with CEO Mark Zuckerberg’s broader vision of developing advanced AI agents capable of handling increasingly complex workflows. However, while the initiative could accelerate Meta’s AI ambitions, it has also sparked a debate about employee privacy, data collection, and transparency.
Credits: AI CERTs
Turning Everyday Work Into AI Training Data
According to reports, the monitoring system has been installed on devices used by Meta employees in the United States. Rather than focusing on a single task or application, the software reportedly gathers information about how employees interact with their computers throughout the workday.
For AI researchers, such data is incredibly valuable. Understanding how people switch between applications, manage workflows, and solve problems can help train AI agents to perform similar tasks autonomously.
The approach reflects a growing trend across the artificial intelligence industry. Companies are increasingly seeking access to real-world behavioural data to make AI systems more capable, context-aware, and effective in practical workplace settings.
If successful, Meta’s AI agents could potentially navigate software interfaces, complete multi-step assignments, and assist users with complex digital tasks without requiring constant human guidance.
Employee Concerns Surface After Rollout
Despite the project’s technological promise, reports suggest that some employees quickly raised concerns after the system was deployed.
One issue reportedly involved the volume of data being transmitted. Some workers claimed the monitoring software generated significant internet traffic, consuming large portions of their home internet data allowances within a relatively short period.
Beyond technical concerns, employees also questioned the extent of the information being collected. Many reportedly sought clarity on what data was being captured, how it would be stored, and whether it would ultimately be used to train AI systems that could potentially automate parts of their own jobs.
The concerns highlight a broader challenge facing technology companies: balancing innovation with employee trust.
Cross-Border Privacy Questions Emerge
Another issue drawing attention is the programme’s potential impact on employees outside the United States.
According to reports, Meta acknowledged that emails and direct messages sent to US-based employees could be captured by the monitoring system, regardless of where the sender is located. This means that communications involving employees from other countries could potentially become part of the collected dataset.
The situation raises questions about international data protection laws and privacy regulations, particularly as governments around the world continue to strengthen rules governing personal and workplace data.
While the project is officially focused on US employees, its indirect reach may extend far beyond American borders.
Meta Defends the Initiative
Meta has reportedly pushed back against concerns, stating that the software is intended to analyse how people interact with computers rather than monitor the specific content displayed on screens.
Company spokesperson Dave Arnold reportedly said that the programme was deployed only on devices used by US employees and that privacy considerations were evaluated throughout both the development and implementation process.
Meta also stated that non-US employees were informed that the system had been installed on devices used by their US-based colleagues and that the initiative was designed to comply with applicable laws and regulations.
The company maintains that privacy risks were assessed and mitigated before deployment.
Credits: People Matters
The Bigger Picture: Building the AI Workforce of the Future
Meta’s reported monitoring programme reflects the increasingly aggressive strategies technology companies are adopting to gain an advantage in the AI race.
The company’s long-term objective is to create AI agents capable of understanding workflows, navigating digital environments, and executing tasks independently. Access to real-world workplace interactions could provide the training data necessary to make those systems significantly more capable.
However, the initiative also underscores a growing tension within the AI industry. As companies seek more data to improve AI performance, concerns around privacy, transparency, and employee rights are becoming harder to ignore.
Whether Meta’s approach becomes an industry standard or a cautionary tale may depend on how effectively it balances innovation with the trust of the people whose behaviour is helping shape the future of artificial intelligence.
