Image: Namibia Fact Check / WhatsApp
A misleading message circulating widely warns WhatsApp users to enable ‘advanced chat privacy’ to safeguard privacy while using the platform.
UPDATE: This article was updated on 25 September 2025 to insert a reference and link to an Africa Check article.
The message, which has been circulating in various forms and posts since about July 2025, and possibly earlier, and was forwarded to Namibia Fact Check in early September 2025, states:
“As of today, AI is available on WhatsApp and has access to all chats” and states that without user intervention, AI, probably Meta AI, “can open group messages, see phone numbers, and even retrieve personal information from your phone … even in private chats.” (See image below for full text)
The Facts
The viral message, or slightly different versions of it, has been circulating in online messaging spaces for a while and has already been debunked by various fact checking entities, such as Indian fact checker Factly in July 2025 and US fact checking service Snopes in August 2025, since appearing in WhatsApp spaces around the world. Africa Check also published an informative fact check of a version of the message on 5 August 2025. Other online media have also debunked a similar message (see here).
In fact, on 19 August 2025, WhatsApp itself responded to another viral version of the claim on X, stating that “this is not true”. (See image below)
The false claim about AI agents or applications gaining unprompted access to private WhatsApp messages and group chats, as well as private information and contacts, comes in the wake of WhatsApp introducing AI features on the platform since early in 2025, with users now being able to interact with Meta AI in chats.
With regard to the call to enable “Advanced chat privacy”, in its July 2025 article Factly summarised it as follows: “The Advanced Chat Privacy feature restricts the use of Meta AI by preventing others from using your messages with AI features, exporting chats, or auto-downloading media, but it does not control Meta AI’s access to chats or contacts, which are already protected.”
And WhatsApp has repeatedly stressed via its social media feeds over the last few months, as it does in the response referred to above, that: “Your personal messages on WhatsApp are end-to-end encrypted by default, meaning only you and the people you chat with can read or share them.”
Our Verdict
Based on the available evidence and statements, the claim that AI agents or applications are a privacy risk because they can independently and unprompted open and access private WhatsApp chats, information and contacts, is clearly FALSE.
To understand how Meta AI interacts with users on WhatsApp, visit the WhatsApp Help Centre on ‘How to chat with Meta AI‘.