Portuguese 
English 
Spanish 
4 min of reading 
0 comments 
With New Security System, Mark Zuckerberg’s Meta Aims to Integrate Generative AI into WhatsApp Without Compromising Message Privacy or User Encryption.
Mark Zuckerberg’s Meta is advancing the use of artificial intelligence within WhatsApp. During its conference focused on generative AI, the company presented a feature called “Private Processing“.
This innovation promises to unlock AI functions in the messaging app without compromising encryption or user privacy.
The proposal comes at a time when the use of AI is expanding but facing skepticism from the public and digital security experts.
-
Artificial intelligence is skyrocketing energy consumption, raising emissions from tech giants and pushing Google, Microsoft, and Meta closer to natural gas.
-
Mercor paid $1.5 million per day for doctors, lawyers, and former Goldman Sachs bankers to teach artificial intelligence to do their jobs, and in 17 months, it went from zero to $500 million in annual revenue while its own contractors accelerated the replacement of their own work.
-
Casio Unveils Moflin, Robotic Pet With Artificial Intelligence Designed To Provide Emotional Comfort And Simulate A Permanent Affectionate Bond
-
Google Maps Revolutionizes GPS Navigation With Gemini Artificial Intelligence, Ask Maps Feature, and Advanced 3D Visualization for Route and Travel Planning
Meta, aware of image risks, bets on an approach that merges innovation with data protection.
What Is Private Processing
According to the company, Private Processing will be an optional feature.
In other words, the user will decide whether or not to activate the features.
Among the expected possibilities are the ability to summarize unread messages or receive suggestions to improve the drafting of sent texts.
These tools will use generative AI, similar to the technology behind Meta AI.
The difference, according to the company, is that this will happen within a secure infrastructure specifically designed to protect the content of messages exchanged on WhatsApp.
The foundation of this technology lies in the so-called Trusted Execution Environment (TEE). This environment would allow user data to be processed in isolation, even in the cloud, without being accessed by third parties— not even by Meta or WhatsApp itself.
Model Similar to Apple’s
Meta’s initiative resembles technology already used by Apple. Called “Private Cloud Computing”, Apple’s model allows AI commands to be processed without exposing user data. The difference is that now Meta is trying to apply this concept within WhatsApp, an app known for its end-to-end encryption.
In the company’s engineering blog, Meta states that the new system will allow, for example, an AI to summarize a conversation without the content needing to leave the secure environment. The same applies to writing suggestions or automated interactions with virtual assistants.
According to the company, privacy will remain the central pillar of the service, and no information will be accessed outside the channel between the interlocutors.
Response to Community Skepticism
Meta acknowledges that this change may generate distrust. After all, WhatsApp is frequently the target of criticism and hacking attempts.
Therefore, the company promises that the new technology will be auditable by independent researchers.
It will also be part of the bug bounty program, which rewards those who find security flaws in Meta’s services.
Despite the announcement, the company did not reveal exactly when the Private Processing will be available. What has been presented so far is merely a “first look,” as the company itself described it.
Zuckerberg and AI That Writes Code
During the same period of announcements, CEO Mark Zuckerberg made statements about the future of software development at the company. In an interview with podcaster Dwarkesh Patel, he stated that he expects, in the next 12 to 18 months, most of Meta’s code to be written by AI agents.
Zuckerberg believes that these systems will be capable of setting goals, conducting tests, identifying errors, and producing code with quality above the average of good human programmers.
He further explained that it is not just about simple automated suggestions. The plan would be to have AI act almost autonomously in software production, replacing a significant part of human labor.
Changes in Discourse in a Few Months
However, Zuckerberg’s statement contrasts with remarks made by him just a few months ago. In a previous appearance on Joe Rogan’s show, the executive predicted that this type of AI would already be functioning by 2025.
Now, Zuckerberg himself admits that broader adoption is only expected to occur between 2026 and the following years.
This change in tone reinforces a point raised by critics of the industry: grand promises about AI often do not materialize. Even when technologies advance, practical implementation is usually slower than publicly announced.
Ambitious Promise, Uncertain Launch
Although Meta has not provided a date for launching Private Processing, the announcement marks a clear attempt to convince the public that artificial intelligence can be compatible with privacy.
The company relies on cutting-edge technology and trust mechanisms to overcome user and expert resistance. Still, the effectiveness and acceptance of the new feature will depend not only on the technology but also on transparency and how it will be communicated.
Meanwhile, the debate about the extent to which artificial intelligence can replace human professionals — or operate ethically — continues. What has been seen so far is a promise: powerful AI, but with guaranteed privacy. Now, the challenge is to fulfill it.
Meta has also signaled that the technology developed for WhatsApp may, in the future, be applied to other products of the company. According to the company, Private Processing may pave the way for new AI uses focused on security and privacy across the Meta platform.
Seja o primeiro a reagir!