In response to the criticisms regarding the integration of AI into WhatsApp, Meta has unveiled a new technology called “Private Processing.” This innovation mirrors Apple’s approach with its Private Cloud Compute, aiming to offer advanced AI features without compromising the privacy of messages. Essentially, this means opting for local processing.
Amazon co-founder MacKenzie Scott has donated over $19 billion to charity in just five years
Diamond batteries powered by nuclear waste promise 28,000 years of clean energy
A Direct Response to Criticism of Meta AI
Since the introduction of Meta AI in WhatsApp, numerous users have voiced their discontent, criticizing it as an unwanted intrusion. The new Meta AI logo is now prominently displayed on chat screens, and the options “Ask Meta AI or search” in the search bar are currently non-removable, leading to significant public backlash.
In this environment, the addition of features like automatic conversation summarization—which requires reading your messages—has intensified privacy concerns and worries about AI training.
A Hint of Apple Flavor
To address these concerns, Meta has introduced “Private Processing,” a system based on a secure cloud environment. The concept is straightforward: data is processed in a secure cloud environment, but neither WhatsApp nor Meta have access to it.
The system ultimately relies on two pillars reminiscent of Apple’s core principles: processing without data retention—even temporarily—and oversight by external audit.
With this strategy, Meta essentially adopts Apple’s major principles: local processing where feasible, otherwise using a secure, encrypted, and auditable cloud, and importantly, no data storage.
In essence, Meta is following the path Apple has paved, making AI more transparent and palatable in apps that are primarily focused on privacy, like WhatsApp, which is still encrypted end-to-end by default. The company has even named its framework “Trusted Execution Environment (TEE),” highlighting a commitment to maintaining a trust-based relationship between the end-user and the service. This is no small detail.