The next time you ask Meta AI about your relationship, job or anything confidential, you may want to check your privacy settings first.
Meta rolled out its AI app in April, taking aim at OpenAI’s ChatGPT. Users can interact with the technology to generate images and videos. Like Instagram and Facebook, the chatbot platform offers users a “discover” feed where you can check out other people’s queries.
Sounds interesting, right? Well the platform is raising some eyebrows this week as some users have been reporting that their most personal prompts and conversations with the chatbot have been publicly posted.
I spoke to Meta at length today about the issue. A spokesperson reiterated to me that all user chats with Meta AI are private by default. They’re not shared in the feed unless you go through a four-step process which includes clicking “share.”
Sounds to me like users may be confused and accidentally posting without realizing that their prompt is now public. Some easy steps to help you advert this problem:
Click on your profile photo
Go to data and privacy
Go to manage your information
Tap make all your prompts visible to only you -> apply to all
Similar to Venmo, I asked Meta why even have a discover feed. Meta told me they want to use the feed to inspire people and show off what others are doing on the platform, whether that means asking for recipes or creating the next great work of art.
I spent the morning on the platform to see how someone might inadvertently post their tax records, legal questions or even a confession. To me, the feed looked like any other AI app: a lot of generated images showcasing political figures, picturesque landscapes and imaginative characters.
Some were asking the chatbot for advice on their upcoming job interview - nothing too personal or embarrassing. But after scrolling for about 30 minutes I did notice something that was definitely not meant to be seen: an apology note to a girlfriend.
These days sticks and stones are mostly just 1s and 0s.
Y
Share this post