Hi there, I am CEO of Mithril Security and we have launched BlindChat, an open-source and privacy-first alternative to ChatGPT that can run in your browser (https://chat.mithrilsecurity.io/) I think privacy issues are a feature not a bug of LLMs as they are literally trained to learn by heart their dataset. Therefore, there is a high chance that if you send data to ChatGPT or such, they will learn by heart your data and someone else might make it spit your sensitive data. You can find more on the topic here: https://dev.to/mithrilsecurity/d... If you are interested by this topic, would love to chat more!