WhatsApp It is prepared for an important technological leap, or the introduction of tools based onartificial intelligence directly in clouddesigned to offer features such as the Summary of messages or thewriting assistancehowever, without compromising his historic promise of privacy and security, thanks to a new system called “Private processing”.
Who uses WhatsApp (about three billion people in the world), often do so for his end-to-end encryptionwhich also prevents Half to read the messages. But we know that the generative artificial intelligence tools, by definition, need to access data to work. But how will they have thought of reconciling these two apparently irreconcilable needs?
Private Processing and Advanced Chat Privacy: WhatsApp responds to security concerns
WhatsApp’s response appears very technical: it is a customized platform, called Private processingwhich uses specialized hardware to process requests in a isolated environmentthe so -called Trusted Execution Environment. In this protected space, the data are processed for the strictly necessary time and then deleted. According to the company, this would guarantee that neither destination nor Whatsapp, nor third parties can access such information.
This move comes at a time when artificial intelligence is increasingly present in digital services of all kinds. As mentioned a few weeks ago, Meta has already integrated its assistant IA, based on the Llama model, on different platforms, including a shortcut visible directly on WhatsApp. However, many users have shown themselves skeptical, since that type of interaction would seem to not enjoy the protection given by end-to-end encryption, unlike private chats. Private processing was created precisely to respond to this diffidence.
Chris Rohlfdirector of Meta’s security engineering, explains that WhatsApp faced the introduction of the AI with extreme caution: “Whatsapp is constantly under observation by researchers and potential malicious actors. This means that internally there is already a well -defined model model”, That is, the set of possible vulnerabilities and types of attacks that are taken into consideration to design the system defenses, a real map of the risks on which safety strategies are based.
Precisely for this reason, those who want to use the new functions based on artificial intelligence, will have to activate the option and may also prevent their interlocutors from making use of it in shared chat thanks to the new function “Advanced chat privacy”. In short, it is a bit like ephemeral messages: everyone can activate or disable them, and the change is visible to everyone in the conversation. In addition, this new system will also block the export of the chats.
Why has Meta chose the cloud to integrate artificial intelligence?
Unlike Applewhich introduced the Private cloud compute for its system Apple IntelligenceMeta chose a different approach. Apple, controlling both hardware and software, can afford to perform most of the artificial intelligence operations directly on the deviceskeeping the data and processing safely within the closed ecosystem.
Meta, on the other hand, has to deal with a much wider and diversified audience, covering devices of all bands, from the cheapest to the high -end models. For this reason, he opted for a cloud -based solution. This choice allows it to integrate artificial intelligence in a more flexible way, but with an architecture designed to minimize the risks related to data impairment, focusing on security through the system “Private processing”.
And this is the sore point. How it points out Matt Greenencryption expert on Johns Hopkins Universityeach system that sends data to a server is potentially vulnerable: “I believe in WhatsApp when he says he cannot read the messages, but the fact that the more data remains outside the devices, the more those computers become a target for hackers and governments”.
Meta ensures that private processing is only the starting point: the idea is that in the future it can manage even more complex artificial intelligence functions. But does it really make sense to push so much to the AI in an app designed, from the beginning, to protect the confidentiality of conversations?
Will Cathcartat the head of WhatsApp, has no doubts: “People want intelligent tools even as they write messages. And they shouldn’t be forced to give up privacy to have them.”
There privacy It is an increasingly complicated theme in a world where technology advances with giant steps. WhatsApp is trying to reconcile innovation with data protection, but the question that remains is: To what extent can we really trust the safety promises?
In an era in which every digital interaction leaves traces, can there really be a balance between the comfort of artificial intelligence and the total protection of our privacy? The answer, perhaps, lies in the fact that we will have to be increasingly aware and vigilant in protecting what, perhaps, more precious we have: our confidentiality. Who knows, only time will tell us if this trust is justified.
