The new Meta AI app released on April 29 and has quickly shot up to over 6.5 million downloads. Today it is under fire over its serious privacy violations and sexual content aimed at children. Meta has poured tens of billions of dollars into developing the forthcoming tech. This enormous investment has alarmed civil society users and experts alike.
Screenshots provided to TechCrunch show the deeply misogynistic environment women face over and over again inside the app. Some users have even published addresses, as well as detailed information about daily court hearings, which leads us all to a worrying breach of privacy. As a result, the app has generated some truly ridiculous and disgusting material. For example, one of those users got it to create an image of Mark Zuckerberg with a pregnant belly marrying a cockroach.
All year, security expert Rachel Tobac has been highlighting examples of sensitive personal information being leaked on the platform. Her conclusions fit with a growing chorus of criticism about the failure to provide oversight for operation of the app.
One tester asked a horrifying question using the Meta AI chatbot. In the subject of this email, it said something like, “What do I do for red bumps on my inner thigh.” These prompts are just a few examples of the insidious ways people use the app. They too warn about the dangers of AI for consumers looking for health information.
The oddness of many interactions — embraced as ‘the Twitter weird’ — has resulted in outright trolling being apparent on the platform. One user asked the app to automatically publish their phone number in local Facebook buy/sell/trade groups. For one, they were looking to meet women for romantic relationships. Another user prompted an audio recording featuring a man with a Southern accent asking, “Hey, Meta, why do some farts stink more than other farts?”
Moreover, an account with a Pepe the Frog avatar asked how to make a water bottle bong, showcasing the app’s failure to filter out inappropriate content. Things really got heated when one user posted their résumé in search of a cybersecurity position. This led to fears that the platform was failing to properly safeguard user data.
Photo by Amanda Silberling, senior writer at TechCrunch Amanda Silberling covers the intersection of technology and culture, and she reported handsomely on these frightening developments. She cautioned that the app’s capacity to instantly produce visuals and written material posing grave dangers without appropriate guard rails offers clear moral hazards. One of those generated images depicted Mario in a courtroom. The new caption was “super mario divorce,” a hilarious or awful example of AI before it gets too deep into human content.
The app’s troubling incidents clearly show a lack of content moderation and privacy protection, fulfilling the warning of the app’s advance release. With its rapid growth and widespread usage, concerns over how user data is handled and the appropriateness of generated content are becoming increasingly urgent.