WhatsApp rolls out a new AI tool with which users can change the tone of their messages, without handing in privacy. The function is called Writing Help and works through so -called private processing.
Writing Help is intended to help users adjust the tone of their messages – from professional to humorous or supportive – without the content becoming readable for Meta or WhatsApp. Thanks to the underlying technology, private processing, the privacy of users remains guaranteed according to the company: neither the original text nor the AI ​​sugots are uploaded or stored by Meta or WhatsApp.
Users activate Writing Help by tapping a pencil icon when they prepare a message in a one-to-one chat or group conversation. This makes suggestions appear in various styles such as ‘Professional’, ‘Funny’ or ‘Supporting’. WhatsApp illustrated the operation with an example sentence: “Please don’t leave dirty socks on the sofa.” The AI ​​then suggested alternatives such as: “Please don’t make the sofa a sofa graveyard” or “Breaking News: Socks Found Chilling on the Couch. Please move them.”
Not yet in the Netherlands
The availability of Writing Help is currently limited. The function is rolled out in phases in English, starting in the United States and some other countries; Expansion to more regions and languages ​​will follow later this year. The function is standard; Users must activate this explicitly in the app settings.
Meta emphasizes privacy-friendliness: the AI ​​processing takes place via Private Processing, where texts are processed anonymously and encrypted. According to Meta, researchers from NCC Group and Trail or Bits, among others, have carried out audits that confirm the safety of this architecture. WhatsApp also emphasizes that the suggestions do not leave any traces after shipment.
By integrating AI aids directly into WhatsApp, Meta wants to prevent users from resorting to external platforms such as Chatgpt.