AI Chat App Exposes 300 Million Private Conversations Due to Security Flaw A popular mobile app, Chat & Ask AI with over 50 million users across Google Play and Apple’s App Store has exposed hundreds of millions of private chatbot conversations due to a misconfigured backend database. Independent security researcher "Harry" discovered the vulnerability, which allowed unauthenticated access to roughly 300 million messages tied to 25 million users. The exposed data included full chat histories, timestamps, user-configured AI models, and custom chatbot names, revealing deeply personal and often sensitive interactions. Users had treated the app like a private journal or therapist, sharing distressing requests such as discussions on self-harm, illegal activities, and hacking under the assumption of confidentiality. The breach stemmed from a misconfigured Google Firebase database, a common but preventable issue in mobile app development. While Chat & Ask AI itself doesn’t operate the underlying AI models (it integrates with OpenAI, Anthropic, and Google’s systems), it handled data storage where the security failure occurred. The app’s publisher, Codeway, did not respond to requests for comment. For users, the incident underscores the risks of assuming AI chats remain private. Even without attached names, exposed conversations can reveal mental health struggles, illegal behavior, or personal secrets, which attackers could exploit indefinitely. The breach highlights a broader trend: as AI tools proliferate, security practices often lag behind adoption, leaving sensitive data vulnerable to simple misconfigurations. No timeline for remediation has been provided, and the full scope of the exposure including whether malicious actors accessed the data remains unclear.