Meta Halts AI Training Program Following Data Leak Crisis

Ender Öztürk

June 25, 2026

Meta officially suspended its internal Model Capability Initiative on June 22 after a significant data leak exposed sensitive information collected through its AI training program. Launched in April, this internal project aimed to track employee keyboard strokes and mouse movements to refine AI models, but it inadvertently allowed widespread access to confidential data across the company. The incident has triggered a severe internal backlash regarding Meta’s data security protocols and surveillance practices. While the company has classified the breach as a SEV 2 incident, the failure to protect employee performance data and personal conversations has raised critical questions about the tech giant’s internal transparency and ethical standards.

The inadvertent exposure of private communications has severely undermined trust within the organization.

Internal Security Protocols Face Intense Scrutiny

The Model Capability Initiative was designed to analyze work processes to improve artificial intelligence capabilities. However, the system’s architecture failed to restrict access as intended, making detailed logs of keystrokes, mouse activity, and performance data accessible to a broader range of personnel than authorized. This lapse in security control has drawn harsh criticism from the staff, who were already skeptical about the invasive nature of the data collection process.

{{WP_IMAGE_1}}

A spokesperson for Meta confirmed that the program was developed with privacy safeguards in place, but noted that all activities are halted until a comprehensive internal investigation concludes. Currently, the company maintains that there is no evidence suggesting the leaked data has been accessed by malicious external actors. Despite these assurances, the incident has highlighted systemic vulnerabilities in how large technology firms handle sensitive internal information.

Employee Morale Declines Amid Privacy Concerns

The pushback from staff members has been swift and vocal, with many employees utilizing internal communication channels to demand greater accountability. The mandatory nature of the program, combined with its subsequent failure, has exacerbated existing tensions within the company. Many workers argue that the collection of granular behavioral data is an overreach, especially given the company’s recent history of significant workforce reductions. The paradox of using data from employees to train AI systems that could eventually automate their own roles remains a central point of contention.

The reliance on employee data for AI development highlights a growing conflict between corporate strategy and individual privacy rights.

Future of Data Collection Remains Highly Uncertain

Meta has yet to announce whether the Model Capability Initiative will resume or if it will undergo a fundamental redesign. The crisis serves as a stark reminder of the complexities inherent in monitoring internal processes for technological advancement. As Meta navigates this incident, the balance between innovation and the protection of employee privacy remains at the forefront of its corporate challenges. The outcome of the ongoing investigation will likely dictate future policies regarding internal data usage and monitoring.

We are interested to hear your perspective on the balance between corporate AI development and employee privacy; please share your thoughts in the comments section below.