Meta has temporarily halted a program that was gathering employee activity data to train artificial intelligence models, following privacy concerns and opposition from its workforce. The initiative, called the Model Capability Initiative, tracked actions such as keyboard usage, mouse movements, and content displayed on computer screens.
The program drew criticism from employees who were worried about the implications of workplace activity data collection, particularly regarding privacy, consent, and trust issues. This led to more than 1,600 workers signing a petition demanding that the company cease the collection of employee computer usage data.
Meta defended the program by stating that it was developed with privacy safeguards in place and asserted there was no evidence of any improper access to the information collected. Nevertheless, the company decided to pause the initiative to conduct a review of the system and its data handling practices.
This development occurs as Meta continues to make significant investments in artificial intelligence, utilizing extensive computing power and resources to enhance its AI models. The company believes that data from skilled employees can play a crucial role in improving the capabilities of its AI systems.
The situation underscores the ongoing debates surrounding the methods companies use to collect data for AI development and the challenge of balancing technological advancements with the protection of employee privacy.