Meta stopped its artificial intelligence training program, which was run within the company and tracked employees’ keyboard movements, after sensitive information was leaked. This development, which emerged on June 22, brought about significant reactions to data security practices within the institution.

The program, called Model Capability Initiative, was launched in April with the aim of developing artificial intelligence models by analyzing employees’ work processes. However, it turned out that the system had wide access throughout the company, contrary to what was intended.

Data security breach worries employees

The data collected within the scope of the program included private conversations, performance records and various activity transcripts. Meta classified this data leak as an internal SEV 2 level incident.

The company spokesperson stated that the program was designed with confidentiality measures, but the activities were stopped until the investigation process was completed. The spokesperson stated that there is currently no evidence that the information was accessed by individuals with malicious intentions.

From the very beginning, employees were reluctant to record detailed data such as keyboard strokes and mouse movements. The revelation that the information was not properly protected and was widely accessible within the company created great rest among employees.

Some employees harshly criticized the deficiencies in information security through internal link channels. The necessity of the program and the resulting security vulnerability negatively affected the working environment within the company.

Meta’s artificial intelligence strategy and employee reactions

Meta recently laid off thousands of employees in order to invest in artificial intelligence infrastructure. The basis of the reactions is that the company collects data from its own workers to train systems that can automate the work of employees.

While the company management argues that this data is critical for developing artificial intelligence models, employees complain that the security of their information cannot be ensured. The situation in question once again revealed the difficulties experienced by technology giants in collecting information.

It remains unclear whether the company will restart this program or what path it will follow regarding data security.

What do you think about this information leak and confidentiality discussions?