Navigating The Fine Print: Using AI to Make Sense of Terms and Conditions
This is a transcript of episode 97 of the Let’s Get Data-Driven podcast.
I’m Lanie Lamarre and I keep my Sephora app cart loaded until I see the “right” samples show up because unless I’m getting the “good stuff” as far as samples go, I’m not buying. Also? I’m not above splitting an order in 2 to get more samples but the trick is that you have to put the order in the next day, and those samples may not be available. I may not play the slots but boss, that’s how I gamble – on Sephora samples!
Today, we’re talking about another semi-compulsive thing in my life: my use of ChatGPT and AI. I’m fascinated by this stuff and wanted to share a use case – and a heck-no-it-isn’t-a-use case – for using it to better understand some more complicated aspects of data privacy.
A few weeks ago on episode 94, I shared with you a resource that I love and recommend to help you better understand the terms and conditions you’re agreeing to when you use a platform, and what the data and privacy implications are for you and the information you share and/or collect on its behalf.
In a perfect and totally responsible world, we would all be reading through the terms and conditions, right? But they can be lengthy and complex to understand where even if you do read them, you don’t always understand what it is you’re agreeing to. Of course, I always encourage you to consult with a legal professional if and when you have doubts or questions about any of this, and anything I tell you here is not legal advice.
At the same time, you also have tools at your disposal to help you better understand these types of documents on your own ahead of seeking out any additional assistance you may want or need.
I’ve been playing with ChatGPT a lot, and one of the things I’ve been testing is to use it to break down agreements set forth by platforms. There are two ways of doing this: you can either copy and paste the terms and conditions, and this would get you the most up-to-date feedback, or you can ask it to help you better understand the Terms and Conditions of a platform.
For instance, I asked ChatGPT: “Can you help me better understand the Terms and Conditions of YouTube?”
And in its enthusiastic and polite way, it informed me that I should really be reading these myself – very responsible, thank you, ChatGPT – and then it went into summaries of its content guidelines, copyright infringement, advertising guidelines, privacy and more.
Now, we remember that ChatGPT and frankly all AI at this point in the game is to considered fallible; all AIs have been known to get things wrong and again, you should be doing you own research and verifications to support anything it tells you. But it’s a great starting point and this is how we’re treating it.
An asset of ChatGPT, though, is that it tends to build on its answers in a conversational mode so I kept asking it leading questions like “How does YouTube use, collect and store my data?” and then “what does it do with this data?” and “what am I agreeing to when I embed YouTube videos on my site?”.
In my experience, this line of questions will give you broad, easily digestible answers that will make it easy for you to start questioning further into. I know I typically read through the brief answer it provides and think, “OK, but wait, I want to know more about this part of it” and you’re off to the races.
One of the gold star questions I have is “when it comes to data privacy, what regrets do people have about using or signing up for PLATFORM NAME?” Again, it will give you a brief, broad answer that you can then pick off like a chicken bone to gain a fuller picture of what issues there may be with whatever platform you’re looking at.
For instance, one of the regrets ChatGPT shared with me in this regard for YouTube was “targeted harassment” and it said, and I quote, “Some users may regret using YouTube due to targeted harassment they may have experienced from other users on the platform, which could potentially involve the use of their personal data.” And I’m like “GO ON!”
Except, of course, that’s not clear enough guidance so what I actually asked was “Tell me more about targeted harassment with YouTube. How does the harassment happen and how is it targeted?” To which it responded things related to doxing and false flagging and other things I have a surface understanding of but then kept talking to ChatGPT like it was my friend until it finally said “Enough! You’ve asked too many question, come back in an hour.”
Let’s get clear about something, though: AI isn’t a lawyer and shouldn’t be treated and seen as such. Treat anything it tells you more as a highlighter pen than anything else where it helps draw your attention to certain things you want on your radar, but anything it says can’t and shouldn’t be seen as any type of gospel or truth.
Furthermore, if you’re using something like ChatGPT, the information it provides only accounts for up to 2021. Meanwhile, regulations are changing at a clip-worthy pace, and so are the terms and conditions of platforms you’re using. Again, treat any information it generates as insight rather than information.
Finally, it bears worth mentioning that while you may copy-and-paste something like a Non-Disclosure Agreement someone else is asking you to sign to get a grasp on what it says before you fire it off for your lawyer to review and approve, you don’t want to be inputting any sensitive corporate information or personal data into these platforms. ChatGPT is under all kinds of fire with the lack of transparency they have around where they’re collecting the data and what they’re doing with the data we input into their systems. In fact, in ChatGPT’s Terms and Conditions, it is quite clear about the fact that it makes no promises as to how personal or sensitive information you enter into its systems will be handled and that you simply should not provide that type of data in any query you make.
Nevertheless, I get that it can be tough to wrap your head around some of this stuff – unless you’re a lawyer who specializes in privacy, data or online businesses, this can feel like a whole other world – but I’m a big fan of making small incremental gains and if you learn something or find a way of seeing something in a way that helps you better understand it – even if you don’t totally “get” it in its entirety – you’re still ahead of where you were and that’s a gain, in my book, every time!
Talk soon – baiieee!