AI Blog: ChatGPT Tour Part 1

Welcome back to our AI blog.  Today, we start taking a more in-depth look at one of the AI tools: ChatGPT.

Our previous blog gave an overview of Microsoft Copilot, Google’s Gemini and OpenAI’s ChatGPT.  This time, we start to take a more in-depth look at one such tool.  Let’s introduce ChatGPT.

ChatGPT

As mentioned last time, ChatGPT is OpenAI’s AI chatbot, accessible at https://chatgpt.com/.  Upon visiting this URL, you’ll likely be prompted to sign in if you haven’t already done so:

You will still be able to send prompts to ChatGPT without signing in, but will be unable to attach any files, adjust your settings or change the model that is being used (more on this next time).

Let’s log in to a free account and take a look at some of the options available.  The first thing that you may notice is the sidebar that appears.  As well as the option to 'Explore Generative Pre-trained Transformers' (GPTs, which we’ll take a look at in a later blog), this sidebar contains the chat history, letting you revisit or continue previous conversations with ChatGPT.

You may also notice that files may now be attached to prompts, allowing ChatGPT to look at and analyse data contained within them.

It’s at this point that many people begin being concerned about the privacy and security of any data, files or information that ChatGPT is given access to. 

ChatGPT collects standard data such as account information contact details, IP address and location.  This data is pretty standard for all websites to collect and, if that’s something that you’re concerned about, you may need to avoid most websites that require you creating an account. 

However, by default, ChatGPT does also collect information from your conversations.  This information is reviewed manually and then used to improve the model.  This does mean that if sensitive data were entered into ChatGPT then it could be seen by a human reviewer for ChatGPT or even returned in future conversations to other people!  Now is a good time to consider the account settings and take a look at how this feature can be turned off.

Within the ‘Data controls’ section of the Settings menu, ‘Improve the model for everyone’ can be switched off to stop the content you share with ChatGPT being used to train future versions of the model.  This option is automatically turned off with the Teams and Enterprise plans, allowing data to be kept safe without worrying about individual users having to remember to turn this setting off. 

Do note that while this information is correct at the time of writing, you should check OpenAI’s privacy policy yourself to understand exactly what is being done with your data in case this has been adjusted or differs in your jurisdiction.

Creating a free account has allowed us access to some more powerful tools, such as attaching files and revisiting previous chats, as well as accessing our account settings.  There are still some features that we are locked out of, such as changing the model in use, but that’s another story for another day.

Join us next time when we consider the options for different models and look at the plans available for ChatGPT!