App Store Data Breach: Millions of Users Exposed

Massive Data Leak Exposes Millions of Users of AI-Powered Apps

A new project scanning the Apple App Store has uncovered a widespread data security crisis, revealing that nearly 200 iOS applications – many related to artificial intelligence – are leaking sensitive user details, including names, email addresses, and chat histories. The findings, spearheaded by security research lab CovertLabs and detailed in the “Firehound” project, paint a disturbing picture of lax data protection practices within the rapidly expanding world of mobile apps.

The Firehound project,first highlighted by user @vxunderground on X (formerly Twitter),is systematically indexing apps that expose user data. As one observer put it, the situation represents “the slopocalypse,” a deluge of poorly secured applications threatening user privacy.

OSINT researcher @Harrris0n created Firehound to address this growing concern, undertaking the “daunting task of hunting AI slop” within the Apple ecosystem. As of January 19, 2026, the project has identified 198 iOS apps leaking user information in some capacity, with a staggering 196 of those apps actively exposing data.

Did you know? – The term “slopocalypse” refers to a surge of poorly coded and secured applications, frequently enough rushed to market, creating critically important security vulnerabilities for users.

Leading the list of offenders is the app “Chat & Ask AI,” which has exposed over 406 million records belonging to more than 18 million users. According to @Harrris0n, a security scan revealed a critical vulnerability in the app, exposing the entire chat history – a total of 380 million messages – to potential unauthorized access. “‼️ STOP USING THIS APP PROMPTLY,” @Harrris0n warned in a post on X, accompanied by a screenshot illustrating the vulnerability.

The majority of these apps appear to be leaking data through improperly secured databases or cloud storage, with many listings openly disclosing the underlying data schemas and record counts. While AI-related applications are heavily represented, affected categories also include Education, Entertainment, Graphics & Design, Health & Fitness, Lifestyle, and Social Networking.

Pro tip – Regularly review app permissions on your devices. Limit access to sensitive data like location, contacts, and microphone unless absolutely necessary for the app’s functionality.

Access to the full Firehound scan results is limited. The project requires users to register for an account to request access to restricted datasets and detailed scan results, with priority given to journalists, law enforcement, and security professionals. “Some scan results are highly sensitive. Until we can responsibly review and redact them, we can’t publish everything in full,” the Firehound team explained. Users who sign in will be prompted to submit a request from their dashboard for access.

Despite initial claims by @vxunderground regarding “AI Slop,” the Firehound website and @Harrris0n’s X profile do not directly confirm that the identified apps were developed using AI-assisted tools like “vibe coding.” However, the prevalence of AI-related apps among those leaking data underscores the potential risks associated with the rapid proliferation of AI-powered applications.

This finding serves as a critical reminder for users to exercise caution when selecting and using mobile applications, particularly AI chatb

Reader question – What steps can app developers take to better protect user data and prevent these types of leaks? Share your thoughts!

Why: The Firehound project uncovered a widespread data security crisis within the Apple App Store.
Who: The project was spearheaded by security research lab CovertLabs and OSINT researcher @Harrris0n, with initial findings highlighted by @vxunderground. Millions of users of affected apps are impacted.
What: Nearly 200 iOS applications, many AI-related, are leaking sensitive user data like names, email addresses, and chat histories. The app “Chat & Ask AI” exposed over 406 million records.
**How