Privacy controls in popular AI assistants

Using AI in conjunction with sensitive data may present some challenges. And we are about to dive into how you can mitigate some of these challenges. My name is Ronny. I've been working with companies that lean heavily on AI for over a decade now. In this course, we'll look at examples like the one we're about to dive into using full out assistance as a service, like ChatGPT. Then we'll dive into more advanced solutions and we'll even look at on-premise solutions such as running your large language models on a computer like this DJX Spark by NVIDIA. Let's go ahead and navigate to ChatGPT. And here you're trusting a third party, which is OpenAI, with your data and you're trusting their best practices. There are some precautions you can take when it comes to configuration. Let's take a look. So I'm going to go ahead and open settings here. And if you go to personalization, you may or may not want to enable memory. You can also head over to data controls. And here, you'll wanna make sure that you opt out of improved model for everyone if you're using this with sensitive information or IP. Now, the main concern is that if your entries are used for fine tuning and training, down the line, a model could potentially regurgitate some of this information. So if intellectual property or sensitive data is involved, you wanna make sure you turn improve model for everyone off. You can also turn remote browser data off and make sure you manage how long you retain things for. Also, navigate to security and make sure that you have things like multi-factor authentication enabled. If we head over to Anthropx Cloud, we have similar options we want to configure. So if I go ahead and open the settings under Privacy, I can also help improve Cloud, and I want to make sure I turn this off, also to make sure that my entries are not used to improve the model. I can also disable the use of location metadata if this is an issue for me as far as privacy goes. So, ChatGPT, Cloud, Gemini, and Copilot are going to be your assistant as a service. You have to trust the provider with your data. In exchange, you get convenience. You wanna make sure you and your team configure things for maximum privacy. Now, there are a few different environments you can use that may give you enhanced control of your data sovereignty. We will explore these in this course. We'll also look at what you can do to mitigate issues when trusting a third party with your data.