Meta AI messages on Instagram do not seem to be encrypted.

Before you go pouring your heart out to Billie “Your Rider or Your Dead Sister” played by Kendall Jenner or AI grandfather Brian on Instagram, know that your message may not be private.

Meta AI characters, now in beta, are a collection of characters, some played by celebrities and creators, that users can chat with on Messenger, Instagram and WhatsApp. However, it seems that messages containing these characters on Instagram are not encrypted from end to end.

With the end-to-end encryption off, the option to start AI chat pops up.
Credits: Screenshot: Mashable / Meta

With end-to-end encryption off, the option is no longer available.
Credits: Screenshot: Mashable / Meta

In the Instagram message tab there is a toggle at the top that allows you to enable end-to-end encryption that protects your messages from unwanted eyes, including meta and government. But when this function is enabled, the option to start AI chat will disappear. If you click on the info button (circle icon “i”) in the chat, the “Use end-to-end encryption” option will be grayed out. When you click on it a window will pop up saying “Some people can not use end-to-end encryption yet” then it indicates that you “can not add them” “Which means AI persona – to chat. Literally you have no choice but to chat with one of these individuals via end-to-end encryption on Instagram.

This window appears to indicate that Meta AI messages are not end-to-end encrypted.
Credits: Screenshot: Mashable / Meta

A major privacy concern with the rise of this generation of AI is the large amount of data that is collected – both model training and giving companies a better understanding of their users. Meta already has a bad reputation regarding the use of personal data. There’s the whole Cambridge Analytica scandal, the Facebook case turning private conversations into law enforcement and the way its algorithms use personal data and behavior to make their platform addictive (and in some cases dangerous) just to name a few. Past cases show that meta – or any social media company – to be fair – should not trust your data No.

When you first test the message messaging AI feature in WhatsApp, you are prompted for immediate rejection, saying, “Meta may use your AI message to enhance AI quality, but your personal message is not sent to Meta.” They cannot read and keep encryption from end to end. “

The denial on WhatsApp says the message has ended up being encrypted, but this has not yet been confirmed.
Credits: Screenshot: Mashable / Meta

This indicates that while certain information about your message may be accessed by AI (still not great for privacy), the content of the message is private. But this is not confirmed, especially due to Meta’s ambiguous AI privacy policy, which states that “when you chat with AI, Meta may use the messages you send to it to train AI models to help improve AI.” .

Mashable contacted Meta to confirm that the AI ​​messages on Instagram were not encrypted from end to end, and also to confirm whether the messages were on WhatsApp and Messenger. While we did not hear back before publication we will update this if Meta Reply.

Last spring, OpenAI launched the opt-out feature for ChatGPT, giving users the option to block their data from being used for prototype trains. However, other AI chatbots such as Google Bard and Microsoft Bing do not have opt-in functionality, despite the ability to delete your activity. On the privacy policy page of Meta AI, there are similar options for deleting your data. You can do this by typing: / reset-ai to delete data from each AI chat and typing: / reset-all-ais to delete all chat data on Meta.