Transcript
Previously we showed that we were able to connect an account. Today I will show what happens after you grant the ChatGPT instance and workspace ID. What a customer would do after connecting an instance is to visit the policy page and to create a API data protection policy. You can see here that I've already created one with ChatGPT. In this case, it's a very broad policy. Any conversation within the ChatGPT workspace will apply to this policy. It is looking for, in particular, a DLP profile that looks for US PII. In this case, it is called ChatGPT policy. Looking at some of my previous conversations, I just entered in this information and I can show in the UI what happens based on this information that I've entered. You would first see an alert in the alert section or scope IT. Here there are two kinds of alerts. They are DLP alerts, but they are alerts based on the two different object types. There is the message as well as the file. If you click into the file, you can see that this channel is sensitive information request denied. You have the object itself, which is the PDF, and the policy name, ChatGPT policy. If you click the first message, it is this message. The exposure is based on the owner. The third message as well here is the message that violated the US personally identifiable information. The user will not only look at the alerts. Following the alerts, they will also look at the dashboard. If you click API-enabled protection, you can look at the dashboard and in inventory, you will be able to see all messages and files. What I wanted to show is for ChatGPT, the previous files and messages, and you can filter based on content. You can look at the attachment, which is here in ChatGPT. You can see the sample data and inventory. This is the file owner, is me, and this is when it was uploaded just a few minutes ago. You can also see the messages from ChatGPT here as well. This is the message that I just sent just one minute ago at 6.20. If you go back to the scope IT alerts, you should be able to see the new messages that I entered in just a few minutes ago. In this case, there is both the message that I sent as well as the message that ChatGPT also sent. One last section where you should be able to see is the DLP incident, which is basically here. These are the three incidents where you have the initial message, 6.09, the PDF that I uploaded as well as the most recent message here. This is the specific message and the channel where this was, or the conversation, sensitive information denied, and this is the date of birth address, the DLP rule violation. Thanks for your time.
