While it may seem harmless on the surface, this feature introduces “a new layer of data processing and potential surveillance into WhatsApp,” Jones warns. The assistant is designed to help users find information, generate content, and interact more easily — but it also means more of your in-app behavior and queries are being monitored and processed by Meta’s AI systems, he says.

According to Meta, the assistant does not access or analyse your private messages unless you initiate interaction with it. However, its mere presence within the chat interface is “definitely raising concerns among privacy advocates and users,” says Laudort.

What’s The Risk: How Much Can Meta See — Can It Read My Messages?

Meta emphasizes that personal messages with friends and family are end-to-end encrypted, the gold standard of security that ensures no one can read your messages, including WhatsApp. If you do talk to Meta AI, it doesn’t link your personal WhatsApp account information on Facebook, Instagram, or any other apps provided by Meta, according to the WhatsApp FAQ.

Meanwhile, Meta says it makes any chat with AI visually distinct so it’s clear they’re different to personal chats.

Yet while it is true WhatsApp direct peer-to-peer messaging is end-to-end encrypted, that “doesn’t tell the full story,” says Jones.

Meta cannot read your messages directly due to end-to-end encryption, but it can access metadata — information about who you communicate with, when and for how long, says Kristian Torode, founder and director at Crystaline.

At the same time, when you interact with Meta AI, it is not protected by end-to-end encryption, and the data can be stored, analysed, and used to train Meta’s models. “Any message you share with Meta AI is, effectively, shared with Meta’s servers. This opens a new vector of exposure,” says Jones.

So, while your personal messages may be encrypted, your intent, behaviour, and AI interactions are very much visible to Meta, he warns.

What Are The Possible Privacy Issues?

AI is built on data and Meta’s algorithms are deeply entrenched in personal information. This is often freely given up by users, says Jake Moore, global cybersecurity advisor at ESET. “This information is heavily analysed and can be used to build detailed profiles on users designed to be sold to advertisers. Meta already knows a great deal about their users but adding AI may bolt on more pieces to the jigsaw it is constantly building.”

User tracking could also be an issue. “The AI’s responses could be based on your past interactions with Meta services, raising concerns about how much data is being cross-referenced,” Laudort says.

While Meta currently claims it does not analyse private messages, concerns remain that AI capabilities could expand over time, leading to more intrusive data collection, adds Laudort.

The biggest concerns are data sovereignty and consent, says Jones. “Many users may not fully understand that using Meta AI within WhatsApp voluntarily feeds personal or potentially sensitive data into Meta’s ecosystem, which can be used to refine advertising, profile users, and train future AI systems.”

In a business or professional setting, this could mean accidentally leaking sensitive company data, strategic plans, or client information without realising it, Jones warns.

Is There Any Way To Turn Meta AI off?

No, there isn’t. Meta doesn’t allow users to disable or fully remove the AI integration from WhatsApp.

You can avoid using it, but the AI icon and underlying data infrastructure remain active within the app, says Jones. Meta may still collect interaction data or usage patterns, even if you don’t directly engage with the assistant, he warns. “In short, you can ignore it, but you can’t opt out of its presence.”

“The feature is deeply integrated into the app and cannot be removed or hidden,” says Laudort. “Many users have expressed frustration over this, as they feel the AI was added without proper consent or an opt-out option.”

While you can’t turn it off, WhatsApp says Meta AI is an entirely optional service that you can choose to use. You need to take an action to chat to or invoke Meta AI. In other words, according to Meta, it’s not ‘on’ until you choose to use it.

What you send to Meta may be used to provide you with accurate responses or to improve Meta’s AI models, so don’t send messages to Meta with information you don’t want it to know. You have the option to delete your interactions with Meta AI.

How To Protect Your Data From Meta AI

If you care about privacy, there are a few steps you can take to protect your data from Meta AI.

Firstly, since you can’t turn it off, avoid interacting with Meta AI, Laudort advises. “Do not tap on the blue circle or mention @Meta AI in chats to minimize data collection.”

At the same time, adjust your privacy settings. “Regularly review and tighten your WhatsApp privacy settings to limit how much data is shared,” says Laudort.

To protect your data from Meta’s AI, users should review app permissions, disable cloud backups and use a dedicated, secure communication platform for sensitive information, says Torode.

You can also use other encrypted apps, such as Signal, which don’t include AI integration.

It looks like the WhatsApp blue circle is here to stay and its data collection will only increase as it becomes more capable. With this in mind, be aware your chats are still protected and make sure you don’t share private information with the chatbot.