There are several ways your AI chat conversations can be made public (www.computerworld.com)

A recent survey reveals alarming misunderstandings around the privacy of chatbot conversations, particularly as Character.AI faces bankruptcy. Many users mistakenly believe their interactions are confidential; however, these discussions can be used to train the bots and are subject to legal scrutiny, including potential subpoenas. In light of Character.AI's financial struggles, the fate of user data is uncertain, with the possibility that valuable personal information could end up being sold alongside the company's assets, as its privacy policy is notably vague about data handling. This situation raises significant concerns within the AI/ML community regarding user privacy and data security. The legal gray area surrounding chat data—especially regarding trained models—highlights a broader problem with how AI companies manage and protect user information. As chatbots increasingly populate our digital lives, the differences in privacy policies among services could lead to significant risks for users, as personal information might be inadvertently exposed either through bugs, sharing features, or absorption into training datasets. With ongoing AI market pressures and the potential for similar incidents at other companies, the need for transparency and robust privacy practices is more critical than ever.