Meta is revising its data handling practices for its AI services, signaling a departure from its previous, more expansive data collection policies. The company, which previously maintained broad rights to user data shared with its AI, is now emphasizing privacy as a core feature to attract users who have been hesitant to engage with its tools.

In a recent 6,500-word manifesto, CEO Mark Zuckerberg outlined a vision for the future of Meta's AI, specifically highlighting the introduction of a 'fully private mode.' This feature is intended to ensure that personal agents operate in a way that restricts access to user data, even from Meta itself or other service providers. This pivot is seen as a strategic move to compete in a crowded AI market where data security has become a primary factor for consumers choosing between various models.

While the company has historically been criticized for its data-harvesting practices, this shift suggests an attempt to rebuild trust. By promising that personal interactions with AI agents will remain private, Meta is attempting to lower the barrier to entry for users who prioritize data security. The effectiveness of these promises remains to be seen, as the company transitions from a model of broad data utilization to one that claims to prioritize user confidentiality.