'Watching Really Smart People': Zuckerberg Pauses Meta Spyware After Shocking Leaks of Private Staff Chats

Meta has paused an internal AI-training programme after employee backlash and a privacy scare exposed the extent of data being collected from workers' computers. The programme, designed to help improve the company's artificial intelligence models, tracked employee activity including keystrokes, mouse clicks and on-screen content, prompting concerns about workplace surveillance and trust.

The decision comes at a sensitive moment for Meta. Chief executive Mark Zuckerberg is investing heavily in AI development and has argued that observing how skilled employees work can help train more capable systems. Instead, the initiative has triggered questions about privacy, internal governance and how far the company is willing to go in pursuit of its AI ambitions.

Employee Revolt Over Workplace Monitoring

The programme, known internally as the Model Capability Initiative (MCI), was created to gather real-world computer-use data for AI training.

According to internal materials cited by Wired, the initiative collected information including keystrokes, mouse activity and content displayed on employee screens. While Meta viewed the programme as a way to improve AI performance, many employees saw it as an intrusive form of workplace monitoring.

More than 1,600 workers signed a petition opposing the initiative, arguing that the company should not collect and repurpose employee computer-use data for AI development.

The petition warned that 'collecting and repurposing this kind of data raises serious concerns around privacy, consent, and trust in the workplace.'

The scale of the backlash underscored growing concern among employees about how workplace data could be gathered, stored and ultimately used inside the company.

Data Exposure Intensified Privacy Fears

Employee concerns deepened after reports emerged that some of the collected information had been more widely accessible within Meta than intended.

MCI data stored on corporate laptops could be accessed internally by employees across the company. An internal notice reportedly referred to exposure involving tables containing 'full prompts and transcriptions, private conversations, people and performance data.'

While there is no indication that the information was improperly used, the disclosure raised fresh questions about whether sufficient safeguards had been in place from the start.

The incident transformed what had been an internal debate about workplace monitoring into a broader discussion about privacy, data security and employee trust.

For critics of the programme, the reported exposure appeared to validate concerns that sensitive information could be accessed in ways workers neither expected nor authorised.

Meta Suspends Programme Amid Review

Following the backlash, Meta announced that the programme would be suspended while the company investigates the situation.

'We have carefully designed this program with privacy safeguards and while we have no indication at this time that any data was improperly accessed by Meta employees, we're pausing it while we investigate,' the company said.

The response highlights the difficult balancing act facing Meta as it races to strengthen its AI capabilities while reassuring employees that privacy concerns are being taken seriously.

The pause does not necessarily mean the initiative has been abandoned permanently. However, it suggests the company recognises both the sensitivity of the programme and the reputational risks attached to it.

Zuckerberg's AI Strategy Under Scrutiny

The controversy is particularly awkward because the project aligns closely with Zuckerberg's broader vision for AI development.

According to reports, Zuckerberg has told employees that AI systems can improve by 'watching really smart people do things'. He has argued that observing how Meta engineers work could help train models to perform tasks such as coding more effectively.

That philosophy helps explain why Meta sought detailed workplace activity data in the first place.

At the same time, the backlash illustrates a growing tension across the technology industry. Data that may be valuable for training advanced AI systems can also raise significant concerns when it involves monitoring employees' day-to-day activity.

As AI companies compete for better training data, questions about consent, transparency and privacy are becoming harder to ignore.

Trust May be the Bigger Challenge

The MCI controversy arrives as Meta continues to expand its ambitions beyond traditional social-media platforms and deeper into artificial intelligence.

Yet the backlash surrounding the programme highlights a challenge that extends beyond any single product or experiment.

Zuckerberg wants Meta to move faster in AI and other emerging technologies. The response to MCI, however, demonstrates how quickly innovation efforts can collide with concerns about privacy, transparency and workplace trust.

For Meta, the immediate issue may be a suspended programme. The larger challenge is whether employees and users remain comfortable with how the company gathers and uses data as its AI ambitions continue to grow.