Cyber Cafe Chats: Would you like Personal Data with that request, Sir?

It's kind of funny when you think about it. You hear all of these people talking about the greatness of AI, how it'll "make your tech job completely useless in a few years," but then, all of a sudden, all of that confidential data you were testing ChatGPT against gets out. Oh yes, dear reader, get that cup of coffee ready, because it's time for another

Plumes' Cyber Cafe Chat

Today's Brew is something that I think a lot of people read about, maybe even argue use cases here and there for, but don't really understand:

What are the dangers of using AI?

If you haven't heard about it already, good ol' Samsung has temporarily banned internal use of AI's such as ChatGPT and others until they can create a safer model after proprietary data was leaked. Now, this isn't anything surprising if you read through any open AI's Terms and Regulations, where most will cite something akin the following from popular AI chat site, Character.AI:

You acknowledge and agree that any questions, comments, suggestions, ideas, feedback, or other information about the Website or Services ("Submissions") provided by you to Character AI are non-confidential and Character AI will be entitled to the unrestricted use and dissemination of these Submissions for any purpose, without acknowledgement of or compensation to you.  

Granted, that's just legalese to say that whatever you give the website, it is fully entitled to do with that information whatever it wishes to do. In most cases, the conversations held with your AI companion are used to train the AI to facilitate chats better, but in training the AI, you might end up teaching it something that should've stayed private. Allow me to explain. 

Let's say that I was writing a letter to a friend, and I asked Bard AI, Google's AI suite, to proofread my letter to make sure it sounds good. It does so, tells me I'm such a wonderful friend and gives me the thumbs up. Well, even though I close the website, my letter still lives within Bard somewhere - in fact, it might be the example that comes up when the next user comes up and asks Bard, "Hey! How do I write a letter? Can you give me an example?"

Now mind you, if I just wrote a simple How-Do-Ya-Do, it probably wouldn't be a big problem; you might get something like my bud's name and my name, but other than that, it's nothing confidential. And that in itself is why I'm not up in arms with OH DEAR GOD, STOP USING AI. Trust me, anyone could find out that information needlessly fast if they wanted to find that on you, and I think AI is still a great reference tool to use so long as you're not marketing yourself on it (looking at you AI "Artists.") No, it's when you start including more personal information that something you probably shouldn't read, like perhaps if I wrote, "Hey man, here's that address for BlahBlah over at Y location. They're usually out and about from this time of day to that, so y'know, don't show up them lmao."

 

Why I would write that in a letter is beyond me, but you get my point. No, the reality is, is that anything you feed these AI bots is free game, not only for them but for anyone on the back-end, and that includes Law Enforcement if need be. Similar to the clause above from Character.AI, there are others listed within these AI website's Terms of Services that quite plainly say, "Hey, if we believe you are of harm to yourself or others, we can give these chat logs to authorities." While I don't think any of my readers would participate in something underhanded like that, it's still good to know.

 

It's hard to differentiate between the real life and the internet at times - and with how share happy the world has become now that our phones are social media machines, sometimes we overshare at times. It's alright if you have in the past! This is just a great learning opportunity, after all! But, I still vividly remember a line from my 1st grade typing class advising us: "Don't put anything online that you wouldn't want your favorite family member or best friend seeing." I'm lucky enough that most of my family isn't around for that first bit, but hoo boy would my Sister strangle me out if I posted anything dumb online. Thankfully, she's yet to find my blog :)

 

Jokes aside, while what I presented above is just an example, the reality is is once you submit something to a bot, or any website for that matter, you no longer have control over it. Even secure websites can get breached, so you need to be mindful of what you share. And, as boring as it is, I highly recommend looking through a modern Terms of Service document to see what all these wonderful websites are doing with your information. It's not my job to put a tin-foil hat on you though, nor would I want you to become overly paranoid about using the internet - it's a fantastic tool that has allowed for countless amounts of good in the world, but I do want to make you aware of what some of the dangers can be so you can better protect yourself. 

You're already reading this, so clearly you have great tastes and are doing the right thing in boosting your Cyber Security smarts. Also, you look amazing! Did you do something different with your hair? 

 

And of course, if you have any questions, please feel free to leave a comment below or email me over at PlumesCast@gmail.com; just know I may read your question on the pre-recorded air so I can train the world a bit more :) Also, feel free to share this article and the blog in general! 

 

I'll see you next time for another Cyber Cafe Chat very soon!