WhatsApp: Meta adds its AI that cannot be removed, Europe rumbles

Difficult to ignore it. Just there, at the bottom right of the screen, this colorful button is disturbing. Barely launched, Meta Ai already takes her ease. In the search bar, a prompt: “Ask Meta Ai or search”. Clearly, an AI that intervenes between oneself and its contacts.

The problem is not only the presence of this button: it is its omnipresence. No adjustment to make it disappear. Nothing to deactivate this function, which gives the impression of being constantly observed. For many, it’s a step too much.

On the networks, negative reactions are fuse, and Meta tries to appease the spirits. On its site, the firm specifies that private conversations remain encrypted and safe. But AI reads everything we send to her. And these exchanges are not protected by usual encryption. Worse still: they can be reused to train future models.

So it’s not just a dialog that can be ignored. It is an entry point to an AI that holds everything, without filter. And Meta’s communication, as well oiled as it is, is no longer enough to mask the rising irritation. And to top it all, Joshua Breckman, director of communication at Meta, slides with a touch of irony in an interview with The Standard :: “There is a version of WhatsApp without AI, it is the one where we do not click on the button.”

Brussels was quick to react. On April 8, a deputy challenged the European Commission: this imposed AI, without explicit consent, complies with DMA (Digital Markets Act)? Meta, already targeted in an antitrust survey, could well find itself once again in the viewfinder.

Irony of timing: a few days earlier, the firm proudly announced Llama 4, its new generation of IA models. A flight forward? A barely disguised provocation? Anyway, the tension rises, and the question remains whole: who really belongs to the private space of a messaging screen?

Difficult to ignore it. Just there, at the bottom right of the screen, this colorful button is disturbing. Barely launched, Meta Ai already takes her ease. In the search bar, a prompt: “Ask Meta Ai or search”. Clearly, an AI that intervenes between oneself and its contacts.

The problem is not only the presence of this button: it is its omnipresence. No adjustment to make it disappear. Nothing to deactivate this function, which gives the impression of being constantly observed. For many, it’s a step too much.

On the networks, negative reactions are fuse, and Meta tries to appease the spirits. On its site, the firm specifies that private conversations remain encrypted and safe. But AI reads everything we send to her. And these exchanges are not protected by usual encryption. Worse still: they can be reused to train future models.

So it’s not just a dialog that can be ignored. It is an entry point to an AI that holds everything, without filter. And Meta’s communication, as well oiled as it is, is no longer enough to mask the rising irritation. And to top it all, Joshua Breckman, director of communication at Meta, slides with a touch of irony in an interview with The Standard :: “There is a version of WhatsApp without AI, it is the one where we do not click on the button.”

Brussels was quick to react. On April 8, a deputy challenged the European Commission: this imposed AI, without explicit consent, complies with DMA (Digital Markets Act)? Meta, already targeted in an antitrust survey, could well find itself once again in the viewfinder.

Irony of timing: a few days earlier, the firm proudly announced Llama 4, its new generation of IA models. A flight forward? A barely disguised provocation? Anyway, the tension rises, and the question remains whole: who really belongs to the private space of a messaging screen?