Published 27 days ago • loading... • Updated 20 days ago

Meta AI App Exposes Users' Private Chats in Discover Feed

Meta's AI app, launched April 29, exposes users' AI chat prompts and responses publicly in a feed linked to their social media accounts.
This happened because Meta did not clearly inform users about privacy settings or that their AI interactions would appear publicly, leading to unintentional exposure.
Users have posted sensitive content including personal questions, legal details, and medical issues, some traceable to their Instagram or Facebook handles.
Rachel Tobac emphasized that when users' assumptions about how a tool operates differ from its actual behavior, it can lead to significant issues with both user experience and security.
The app's privacy defaults have sparked concerns of a major privacy failure and user embarrassment, raising questions about Meta's handling of user data and AI transparency.

Insights by Ground AI

Does this summary seem wrong?

79 Articles

All

Left

Center

Right

Corriere Della Sera

Center

Ask Meta AI for the Number of the Railway Company and Gets Him that of a Private Citizen: Nightmare Privacy

According to the Guardian, the address of a private citizen was accidentally shared on WhatsApp with a user who had asked artificial intelligence for help. And the secrets of some chats became inadvertently public domain

20 days ago·Italy

Read Full Article

Hong Kong Free Press

Lean Left

Meta sues Hong Kong-based company behind AI deepfake app that creates fake nude images

US social media giant Meta has sued the Hong Kong-based company behind an app that uses artificial intelligence to create fake nude images of people without their consent, seeking to stop the firm from advertising the app on Meta platforms. Facebook. File photo: Bastian Riccardi, via Pexels. Meta said last Thursday that it had filed the complaint against Joy Timeline HK Limited, the maker of the CrushAI app, amid what it called a “concerning gro…

20 days ago·Hong Kong, Hong Kong

Read Full Article

New York Magazine

Left

Meta’s Privacy Goof Shows How People Really Use AI Chatbots

The Meta.ai public feed offers some uncomfortable insights, writes technology columnist John Herrman.

21 days ago·New York, United States

Read Full Article

BleepingComputer

Center

Asana warns MCP AI feature exposed customer data to other orgs

Work management platform Asana is warning users of its new Model Context Protocol (MCP) feature that a flaw in its implementation potentially led to data exposure from their instances to other users and vice versa. [...]

21 days ago

Read Full Article