WhatsApp It is working on a new functionality that could revolutionize the way we manage the most crowded chats, especially in the groups or channels that are populated by hundreds of messages every day. It is a system of automatic summaries generated by artificial intelligencedesigned to help users recover the context of discussions quickly with a simple “tap” on a dedicated key, without the need to read everything from start to finish.
The novelty, identified in the Version 2.25.15.12 Of WhatsApp Beta for Android from Wabetainfo (a portal specialized in the discovery and dissemination of the latest news present in the beta of the messaging app), will take advantage of a new technical infrastructure called Private processingwhich guarantees the safe and reserved processing of messages. In practice neither Half (the company that develops WhatsApp) nor will others be able to access the content of the conversations. Privacy therefore remains at the center of the project. The summary, generated by Meta Aiwill be completely optional and visible only at the user’s request, which will thus be able to decide when and how to use AI to summarize the content of a chat. It is a functionality designed to simplify digital life, especially in an era in which theinformation – or the excess of information – is on the agenda. At the moment it is not known when the function will land in the stable version of WhatsApp.
How the summaries of the chats work and how the function can be activated on WhatsApp
This new function to generate summaries of the WhatsApp chat with the AI It is designed to offer a synthetic summary of the most recent messages received in any conversation – whether it is a single chat, a group or a channel – as soon as a number of unread messages are accumulated. It will be the same WhatsApp app to propose, through a special Dedicated buttonthe creation of the summary and the user can decide whether to actually generate or not. If so, a destination will generate an automatic summary of the content, allowing to quickly recover the thread of the speech, without having to flow hundreds of notifications or manually seek the most relevant information.
In all this, theprivate processing infrastructurea technical architecture developed with the principle of “privacy by design”. Translated in pennies, it means that data protection is not a secondary addition, but a central feature, integrated from the design. In concrete terms, this system uses encrypted connections, protected execution environments, anonymous identity and client verification to make sure that no one – not even the engineers of Meta or Whatsapp – will be able to read or keep the requests or messages sent to generate the summary. The data will not be saved on external servers, and the entire process will be managed temporarily, directly on the user’s device or within a safe network.
It is important to clarify that theUse of the function of generation of summaries will be optional. It will not be activated automatically, and users will decide whether and when to use it. If you choose to activate it, the summary will be generated only at the moment required and shown in real time, without archiving the original contents. The messages that will be summarized will continue to enjoy the end-to-end encryptionthe security standard that guarantees that only sender and recipient can read the content of a chat.
The usefulness of this function is evident above all in those contexts in which the amount of messages becomes difficult to manage: school groups, professional communities, information channels with continuous updates, and so on. In all these cases, being able to access an automatic summary of recent messages allows you to stay updated without having to read every single intervention, saving time and effort.
Guaranteed privacy: limits of the summaries of chats with the AI
Let’s pass, now, to some limitations of the summaries of the WhatsApp chats made with the AI. First of all, it is good to clarify that the function will not be available in the chats in which you have activated the so -called Privacy Chat advancedan optional mode that blocks the use of functionality to even if the processing system is considered safe. In commenting on the reason for the thing, Wabetainfo explains:
This decision (not to generate the summary with the AI if the advanced privacy chat is active, editor’s note) does not derive from doubts about the private processing capacity to protect requests and messages, but instead respects the preferences of users who have chosen not to use the functionality AI in certain conversations.
From a technical point of view, part of the infrastructure on which it is based Private processing It will be released as Open Source, i.e. made publicly available to be analyzed, tested and improved by independent researchers and scientific communities. It is an important step towards transparency, which allows the community to evaluate the real security guarantees offered by the system and, possibly, contribute to its improvement.
