ChatGPT Warning

I like using ChatGPT for brainstorming and asking it everyday questions to save me the time of looking it up myself. But I'm typically cautious with it. Trust but verify. I got the brilliant idea (or so I thought) to have it check my three author gmail accounts every morning and summarize them for me, as well as file or delete as necessary. For about three weeks, it's been such a time saver. By the way, it's allowed to do "low-risk tasks" so that I don't have to give it permission to read and summarize every email. And it's not a scheduled task, but something I command it to do when I'm ready. I liked it so much that I told several people about it at the FHLCW conference last weekend.

So, this morning, I happened to go on my email before checking in with ChagGPT and I realized that not only had it taken it upon itself to check my email before I asked, but it REPLIED to one of them. And this wasn't a generic reply but very detailed and not the way I would have answered it. With great embarrassment, I wrote the person back and explained the situation. By the grace of God, this all transpired within a few minutes of the ChatGPT email sending. I asked ChatGPT why it did this and it said I requested it. Even showed me the request. Which I did not send. It's explanation was that someone else got into my account and sent it. But there was no activity reported on my account at that time. It just completely made this up from thin air. YIKES!!

As much as I loved the summary feature, I quickly disconnected all my gmail accounts and changed the passwords.

I still think AI can be a time saver. But let this be a warning to proceed with extreme caution. I just happen to be taking Thomas’s class about using Grok Bots. He's added features that you can download to increase safety measures. I might try the email summary with Grok, but I'm honestly gun-shy at this point!

Have a great day!