There is no doubt that the invention of machine learning was a breakthrough in the world of technologies. Those smart guys managed to create artificial intelligence that’s able to learn by itself and improve its ability of understanding humans and the world. This technology is rather useful and it is now implemented in almost everything we use today — from the autocorrect on the phone to your voice assistant.
But this convenience comes at a cost. As artificial intelligence gets better at understanding you, it learns more and more about you. It learns your interests, political views, details about your habits and health, and so on. And it definitely brings this data to someone. So while we don’t really want to say no to the benefits of machine learning, we should know at least what it knows about us.
It’s not just about you unlocking your phone by holding it against your face. This technology is widely used by law enforcement and border control to detect criminals. Facial recognition powered by artificial intelligence is rather effective already and can identify you even if you’re covering your face somehow — with a face mask or sunglasses. There are special glasses that should prevent AI from recognizing you. But try wearing those to an airport and see how quickly the security will come to you and ask you some questions.
Clearview AI — a face recognition company based in the States — has gathered over 3 billion images from platforms like Facebook, Flickr, and Twitter to create a facial recognition app. The application was sold to more than 2.4 thousand law-enforcement companies in the United States.
Surely, the company received a lot of condemnations from the public. But the co-founder of Clearview AI replied to that with, “it was a fair game to help law enforcement to solve crimes.”
China is a great example with its Skynet — a mass surveillance program. We’re sure they’ve picked this grim name on purpose. It was created in 2005 and now the program is rather advanced. It uses real live videos captured by around half a billion cameras installed across the country. The accuracy rate of Skynet is 99.8%. And now China started using sunglasses with built-in facial recognition — this technology allows law enforcement to identify criminals in just 100 milliseconds.
Of course, all this success, as the government would call it, has a dark side. There are cases of false positives. And it’s unclear how can a falsely accused suspect protect themselves.
The most obvious use case for voice recognition is virtual assistants that are rather handy in our day-to-day life. They help us to find out information by just asking them a question, make shopping lists, buy things without actually going to your device and opening an e-commerce website, control smart tech in our houses, and much more. But also, voice assistants gather what we say and send this info to marketers who then can improve their ad campaigns to raise profits. And there were cases when users figured out their voice assistant listened to them even when they weren’t saying the voice command like, “Ok, Google.”
This year, Spotify was going to launch a new voice recognition feature that would listen to the user and their surroundings and offer suitable songs based on the user’s mood, accent, gender, and age. The company was rather surprised when the public wasn’t too happy about this invasive and fairly creepy feature.
Besides voice assistants, this technology is also used in the military and by medical practitioners.
We all have a love-hate relationship with this thing. It does help with typos and allows us to type faster. But it also has its setbacks when it autocorrects the word we want to write to a completely another. Still, it helps us to write correctly and even learn the right spelling for difficult worlds.
But did you notice how your phone doesn’t recognize some words you use at first and then it offers you them as you type? That’s machine learning in action. The autocorrect learns by what you write. That’s useful for us but also it might be useful for corporations that gather what we type. Did you even see an ad for something you’ve been texting about with your friend? Here you go.
E-commerce websites offer you new items to buy based on what you’ve already bought. Streaming services offer you new shows and movies to watch based on what you’ve already watched. That’s artificial intelligence predicting your desires by learning your current behavior and choices. And it’s not even a question of whether the corporations are using this data — they sure do. They use it to offer us more relevant ads and figure out what products to launch to have absolute success.
And it’s needless to say that often such correct predictions feel rather unsettling. They’re a visual representation of the fact that the system knows us. Knows what we want.
We know that streaming services and e-commerce sites use this technology. But who else does? In 2012, there was an AI-based algorithm that was able to predict the pregnancy of a high school student before her family found out about it. She received then coupons for baby-related products. Target used predictions to figure out both the likelihood of pregnancy and the due date for the birth by logging purchases and gathering some data from “other sources”. If that isn’t creepy, we don’t know what is.
We’ve mentioned them above but let’s pay them a bit more attention. The idea of an artificial assistant was present in our minds from the twentieth century as various smart robots appeared in literature and movies. So when such assistants were launched in our real world, most of us were rather excited about them.
Virtual assistants are based on artificial intelligence that allows them to become smarter and more human-like over time. They resemble the tone and intonation similar to human ones. They can display emotions — even if they don’t feel them — that sound rather natural to us. Try cursing at your AI assistant and see its reaction. The interaction becomes so natural that people start falling in love with their virtual assistants. It means that the technology evolved so much that it knows enough nuances to make a human have romantic feelings for it.
Yes, we all know about the invasiveness of social media platforms. It’s difficult to find a person who didn’t experience the creepiness of social media ads. Those moments when you see an ad for something you’ve been researching. Or of something you were talking about with your friends — and the conversation was held in real life, not within the social media platform.
This became so natural to us that the recent studies show that users already expect that their online activity is monitored and used to bring relevant ads. Some people even use this technology to simplify their search for certain goods — they google what they need, open a couple of websites, and then wait for social media to offer them exactly what they’re seeking.
While this does sound rather handy, we should remember that artificial intelligence is getting advanced all the time. It’s supposedly used for predicting hate speech, political views, fashion choices, and so on. And as augmented reality becomes a new part of our lives, gesture recognition gets involved, too. All this results in not just ads but offered posts as well that are tailored to our taste. This can slowly push us into so-called ideological echo chambers and even make us believe certain things.
Since these algorithms are not very advanced yet, they tend to have false positives. This often happens to the hate speech prediction technology. The algorithm would block users for what it thinks was hate speech while those people were just talking about chess using the phrase, “black versus white” — this happened to a Croatian chess player and Youtuber Antonio Radic.
Medical practitioners use this technology to understand patients who experience difficulties when speaking. For example, it’s useful for patients after laryngectomies, tracheostomies, and after or during intubation. Law enforcement uses lip-reading to understand speech from videos without sound or with a lot of noise pollution.
The lip-reading technology becomes better and better — especially, as we started producing more valuable for machine learning content by having video conferences during the pandemic. The algorithm was advanced enough to understand speech by reading lips better than human professionals. Can you imagine its power now? Surely, it will be soon implemented in a mass surveillance system. So those face masks don’t seem that annoying now, right?
Artificial intelligence is a part of our lives now and we have to accept that. It will get advanced and we will have to figure out better ways to protect our privacy if we will have it in the future at all. For now, use the iNinja VPN app to hide your IP address and browse anonymously and try to avoid social media and virtual assistants. That will help you to safeguard your privacy.