As we navigate the world of emerging technologies, including AI chatbots like chatGPT, the question often arises on data security.

By default, OpenAI uses the messages you send to ChatGPT to "train and improve" its models. However, you have the power to protect your privacy by using the platform's incognito mode. In this mode, your conversation log with ChatGPT is deleted from your account when you close the browser tab. It's important to note that OpenAI retains this data for 30 days behind the scenes to monitor for abuse before permanent deletion.

According to ChatGPT's privacy policy, they may collect and log user interactions for improving their models and services. The policy also mentioned the possibility of sharing aggregated data with third parties but did not provide specific details about the identities of these third parties.

To safeguard your privacy, it's wise to limit the sharing of personal information, following the general rule applied to most online platforms. By doing so, you not only protect yourself from potential privacy concerns but also from security breaches. For instance, a recent ChatGPT outage resulted in users viewing titles from other active users' chat history.

Thankfully, there is an alternative solution of a locally hosted Large Language Model (LLM). We call it Local AI Bot, where you are deploying your own LLM on your consumer-grade computer offline, you gain full control over your data and privacy. While the quality and performance of this local LLM may not match that of ChatGPT, it provides a potent alternative that can be customised and trained to suit your requirements.

In the video we compare between a smaller-sized local LLM with 7 billion parameters and ChatGPT with 175 billion parameters.