Artificial intelligence is the topic of the day. Everybody is chatting about it and wondering how it will soon be stitched into the tapestry of our lives. But there’s one AI topic thread that you may not be thinking about or talking about over your morning cup of joe: What is AI borrowing from you and using right now?
Yes, indeedy. We rarely think about how AI uses the information that we didn’t know we were offering. You see, AI tools, such as chatbots and the like, fall into the category of an LLM or a Large Language Model. People creating these hungry, growing LLMs must feed them an incomprehensibly huge and diverse amount of raw data in order to train them up well. And that tasty, tasty stuff comes from all of us: users of the internet and social media.
For instance, OpenAI, a pioneer company in artificial intelligence research, recently revealed that it develops its AI models (including ChatGPT) by tapping into three primary pools of information: data publicly available on the internet, data licensed from third parties, and data provided by users. And it noted that those forms of data gathering are fairly universal.
It’s Free for the Finding
But let’s get back to you and your data. How does your information become a part of that mix?
Well, one of the initial methods employed for data collection is something called “web scraping.” That’s essentially the process of gathering up all the online data that is openly available to anyone using public databases. No big deal, right?
Well, you may be surprised how much there is about you out there. Your name and date of birth, current and past address, your posted family photos and selfies, your employment history. All of that is as easy to grab as using a simple web search. And then there are things about you that were hacked, leaked or grabbed unlawfully. Medical information and private tracking data sometimes falls into that category. A thorough web scrape can pull all those things up, too.
It’s For Sale
Secondly, companies can purchase data sets connected to you. Unless you’re an attorney or someone obsessively detail oriented, you likely do what everybody does when they pick up a new app or sign up for a social media account. You quickly scroll through the terms of service and say … whatever. But with most of those agreements, the fine print allows the company in charge to use or sell your user data as they see fit.
Even banks can, and often do, sell your credit card purchase history (along with other financial information) to third-party companies.
For instance, a friend of mine recently decided to stop and buy something on the spur of the moment. He didn’t call anyone about it or mention his purchase to his wife. He didn’t even speak about the item within “earshot” of his cellphone. But by the time he got home, bammo, there were ads for the very thing he purchased all over his socials.
And It’s Theirs, Even if It’s Technically Yours
Those terms of service terms you agree to also let the companies use your data even when you thought it was proprietary information that you owned. So, things you post on your social media accounts—personal pics, thoughts about politics and social issues, queries you make, websites you browse, things you buy, heartfelt feelings you share or videos you watch—can all be used by the AI of that given site.
Instagram, for instance, boldly states in its terms of use that “Instagram does not claim ownership of any Content that you post on or through the Service.” That seems straightforward enough. But it goes on to say: “Instead, you hereby grant to Instagram a non-exclusive, fully paid and royalty-free, transferable, sub-licensable, worldwide license to use the Content that you post on or through the Service.”
In other words, yep, it’s yours. (And it’s theirs, too.) That kind of verbiage, which secures companies a broad license to use anything users post, is fairly common. Companies can then freely use collected data to train AI models, compile answers, customize AI responses and the like.
Again, that may not amount to a hill of beans to you. You may not care a bit that all those parts of you are being used. But, then again, you might.
How do You Say No Más?
So, is there any way to keep companies from selling your random musings or posted images to third parties you’ve never heard of? Yes. You may not have realized that you were signing up to be a friendly, neighborhood AI contributor, but in some cases, there are ways you can opt out.
CNN posted an article detailing the opt-out process at the end of 2024.
The CNN article suggested that in some cases, such as with Meta, you’re required to jump through a few hoops if you want to shield your data, including making your account private. Others, such as LinkedIn and X offer pathways to navigate to an opt out button. And some, such as Reddit, say, “Sorry, opting out of data usage ain’t an option.” But Reddit does say that private content, “such as private messages, posts in private communities and browsing history,” won’t be shared with third parties.
Whether you can trust these companies as far as you can throw them, I couldn’t say. At the very least, we should be aware of how these things work and how we’re contributing to the data-mining process. We can be thinking about the services and devices we use and looking into the settings we allow.
For example, there is some sort of Privacy and Security section in the settings menu of every phone. And if you’ve ever worried that your phone is secretly listening to you (and, yes, it is) you can usually disable the microphone options in your app settings. You can also check for ways to limit how much data an app or service collects.
There are options to take and choices to make. But the first step is being informed. We at Plugged In are here to help.
