Meta warns users to “avoid sharing of personal or sensitive information” in their AI apps
Meta seems to have finally taken a small step to address the popularity of overclassification happening in public feeds of its AI applications. The company has added a brief disclaimer that warns users to “avoid sharing of personal or sensitive information” in the “fort for Feed” button in the Meta AI application.
Change is first Business Insiderthe app is “one of the most frustrating places to go online” due to the enormous intimacy, awkward, and sometimes personally identifiable information meta-users – apparently unconsciously – publicly sharing the app’s built-in “Discover” feed. Although Meta AI does not share the chat history of users by default, it seems that many app users choose to “share” their interactions without realizing that it makes sound and text chat visible.
Last week, I found a post by users who made suggestions for “improving bowel movements” and asked relatives if they were responsible for the employer’s unpaid taxes. Another user “keeps this private” into his public post, apparently covering up his embarrassing chat afterwards. These types of weird public interactions have been happening since the Meta AI app launched in April, but it caught a renewed focus last week after social media users began posting all the weird conversations that were visible in the app’s “Discover” feed.
Privacy experts criticized Me and pointed out that most other mainstream AI chatbots do not include social, publicly visible feeds. “If users expect how the tool features match reality, you have a huge user experience and security issue,” Rachel Tobac, a security expert who worked with Meta last week. “Humans have built patterns around AI chatbots and don’t want their AI chatbot tips to show up in social media style to discover feffs, not how other tools work.” Mozilla Foundation Meta changes the design of the app. It wrote last week: “It’s not obvious that the Meta AI app is that what you share is completely public.”
Now, the company has obviously noticed it. With this change, choosing a shared meta-AI interaction will publicly prompt the above warning, although it seems to appear only in the first part. “What you are prompted to post is public and can be seen to everyone,” it states. “Meta may recommend that you prompt other Meta applications. Avoid sharing personal or sensitive information.”
As Business Insider Note that the app’s public feeds no longer seem to feature text exchange text that other users share with the app, with only AI-generated images and videos. It is not clear that the recent negative attention the app has received is a permanent change. We have contacted META for more information and will be updated if we hear back.
Meanwhile, if you find yourself a victim of an unexpected public post in the app, you can delete it by clicking on your profile in the upper right corner of the app and head to Data and Privacy -> Manage your information -> Make all public prompts visible to you only And select “For Everyone”.
