What does artificial intelligence know about you? Way more than you think.
You’re probably aware that companies and governments collect data about you, like your name and where you live. But your digital profile isn’t based only on the data you choose to share, all sorts of additional data about you is generated by machine learning algorithms. These algorithms analyze things like what time of day you’re online, whether you’re prone to typos, and what kinds of things you search for.
These algorithms use this data to make decisions about you and for you, and they do it every day – often without us even knowing. It usually happens “behind the screen”, without bothering us, in a smooth and efficient manner.
Algorithmic decision-making can carry more weight than you might expect. While algorithms do innocuous or helpful things like changing the traffic signals when you approach an intersection, they also decide what content to show in your social media feed. There are also algorithms that assist real people in deciding whether you can get a mortgage, get into a particular university or qualify for insurance.
But how do they get information about you to make those decisions in the first place?
It’s all based on your data – and what’s known as your digital profile. Learn more in the video below featuring human and technology rights lawyer Katarzyna Szymielewicz and president of our partner organization, Panoptykon Foundation, or read on.
The three layers of your digital profile
The first layer is the one we control ourselves – it’s information that we often give voluntarily when we use computers, phones, or anything else connected to the internet: your username, your real name, your photos, your friends, your likes, what you search for, who you’ve blocked from your contacts. In most cases, you can choose not to share this type of data through a combination of opting out, avoiding data-hungry platforms and deleting your activity history, but doing so will require constant attention and self-control. It’s a rather difficult task if you simply want to live part of your life online with some degree of spontaneity.
The bad news is that even if you manage to control this first layer of your profile, there are two more that are built without you even knowing. By the mere fact of being online and doing things online, you are sharing a lot more than you might think.
The second layer in your digital profile includes traces of what you do online – the data collected about your data or “metadata”. This could be the sites you visit, your current location, and even stuff like how fast you type (and how many mistakes you make!)
It also includes a stamp of any “connected” activity. For example, if you’ve got a smart fridge, it knows you held the door open for 3 minutes at 2am for that late-night snack. And it won’t forget.
It’s nearly impossible to prevent this type of data being collected about you, or to opt out from this constant behavioral observation.
What else could algorithms learn about us after they’ve already observed every move we’ve made online?
They can learn more by comparing the data they collected about you from layers 1 and 2 with data they’ve collected from others, and establishing patterns in that information. Machines make assumptions and predictions about you based on statistics. They use all the prior “knowledge” gathered about you, as well as observations made about your friends, or people similar to you, to predict things like your ethnicity, who you might vote for, whether you might be expecting a baby, if you love gambling, and even if you suffer from a mental illness.
It is this third layer of your digital profile – called “inferred data” – that is the most valuable for companies using AI. Based on these assumptions, they decide things like which ads they should show you, and whether you should get a loan.
And it’s pretty much impossible for you to control or even discover what these machines think about you.
The good news?
People are starting to do something about this problem so that you can regain control of your information. Our partners at Panoptykon are making sure that there are legal safeguards in place to protect us, and that legal complaints against data-hungry companies result in big fines. Here at Mozilla, we’re rethinking how your data is collected and used altogether in The Data Futures Lab, an experimental space for instigating new approaches to data stewardship which give greater control and agency to people. It’s a long march but we are heading in the right direction.