Sounds like the beginning of a 21st century horror movie. The history of your browser has been public and you didn’t know it. That’s basically what I feel like with the new, standalone meta AI apps today. There, many people are publicly revealing ostensibly private conversations with chatbots.
When you ask the AI questions, you have the option to press the share button. This will take you to a screen that shows you a preview of your post and allow you to publish it. However, some users seem unaware that they are publicly sharing conversations, audio clips, and images of these texts with the world.
When I woke up this morning, I didn’t expect to hear an audio recording of a South Accent man asking, “Hey, meta, why do farts smell more foul than other farts?”
Flatulin-related inquiries are the least common among meta issues. In the Meta AI app, we have seen people seeking tax evasion help if a family member is arrested for proximity to a white-collar crime, or write a character reference letter for an employee facing legal trouble, if the person’s first and last name are included. Others, like security expert Rachel Tabak, have found examples of people’s home addresses and sensitive court details, among other personal information.
When TechCrunch arrived, Meta spokespersons did not comment on the record.
Even if you admit to committing a crime or creating a strange rash, this is a privacy nightmare. Meta does not show users what their privacy settings are or where they are posting when they post. So is the search for how to meet “Big Booty Women” if you’re logged in to Meta AI on Instagram and your Instagram account is public.
If Meta didn’t ship the app using the crazy idea that people want to see conversations with each other’s Meta AI, or if someone in Meta could foresee this kind of feature would be an issue, then much of this could have been avoided. There are reasons why Google has never tried to turn search engines into social media feeds, or why AOL’s pseudonymized user search publication became awful in 2006. Disaster recipes.
According to App Figures, the Meta AI app, the App Intelligence Firm, has only been downloaded 6.5 million times since it debuted on April 29th.
While that may be impressive for indie apps, we’re not talking about first-time developers creating niche games. It is one of the richest companies in the world, sharing technology and apps that have invested billions of dollars.
As each second passes, these seemingly harmless enquiries about meta AI apps approach virus confusion. In a few hours, more and more posts have been featured on apps that show clear trolling, like those who share their resumes and seek cybersecurity jobs, or Pepe’s account, where Pepe asks how to make a bottle of water.
If Meta actually wants people to use that Meta AI app, public embarrassment is certainly one way to get attention.