This is a link to the Playlist for Reading This Article: tracks are linked throughout.
[Track 1: Control start at :40]
Adele is what I call banana music. I don’t hate it or anything, but I gotta be pretty darn hungry for music if I’m gonna listen to Adele. But, I’ve been listening to her new album non-stop for a very different reason altogether.
Because when I listen to Adele I’m telling the machines that persistently watch us all that I am not who I appear to be.
Rolling in the deep
We could have had it all
(You’re gonna wish you never had met me)
Rolling in the deep
(Tears are gonna fall, rolling in the deep)— Adele
Yes. Adele is right. We are in too deep and we are nearing a point where escaping the grip of AI is impossible. We’ve been letting the intelligent machines get incredibly intelligent without ever really considering what we are teaching them.
And now the machines are manipulating you. Algorithms guide your most intimate choices. They feed you information to shift your opinion and deciding what you know and how you enjoy everything you love. Tears are gonna fall, rolling in the deep.
What can you do to maintain a wee bit of agency in a world that is painted with AIs that watch us with cameras and surround us with screens and speakers? It’s actually easier than you think. Because, like any big boss in the side-scrolling video game of life, it appears invulnerable, but there is a chink in the armor of AI and you can exploit it if you know where it is.
The AIs that are used to target you, understand you entirely based on the choices you make online. Therefore, if you choose carefully, you can convince the algorithms that you are an entirely different person.
Creating Identity Camouflage
[Track 2: Fight The Power start at :23]
To manipulate the algorithm, it’s important to understand how AI sees the world. It has such an incredible view of data, that it has a hard time noticing details. Like Andre The Giant fighting Wesley in The Princess Bride, it’s so powerful it doesn’t know how to handle just one person at a time.
In fact, neural networks are so complex that the people who build them don’t actually know what is happening when they “think.” In much the same way the world’s greatest meteorologist prognosticators can really only point at the sky and shrug when a storm is coming, the bajillions of computations within the brains of computers are pretty hard to understand to our feeble human minds. A hyper-intelligent Human still learns one image or line of text at a time. But, machines that run the world are “taught” on gigantic libraries of data like Wikipedia. This makes the machine incredibly powerful, but incredibly myopic. They have complete photographic recall, but like Kim Peek, (the real “Rain Man”), they can’t understand what they know. They can emulate a writer’s style perfectly, without knowing what the words actually mean. They understand people by looking at what “millions of people like you” would do, but they have no idea how to evaluate the nuance and complexity of a real human being.
This means you can make small choices to disguise yourself online and trick AI into thinking you are someone else entirely. And the impact can actually be quite big.
Understanding Your Blast Radius
Don’t underestimate the things that I will do
There’s a fire starting in my heart
Reaching a fever pitch and it’s bring me out the dark.— Adele
The key to the revolution is that machines have complete and total access to your data and unending computational power, but they are limited by their form factor. For example, your AI powered coffee machine might want to murder the president and replace him with a robot, but all it can really do is murder a cup of coffee. So, we aren’t going to fight Big Tech. We’re going to fight for our right to make playlists.
[Track 4: Who Am I? By Snoop Dogg]
Music listening is an area of the web where the machine is looking at very specific behaviors to control you and where you are aware of the choices you’re making. Spotify is a great place to start. It’s one of the most data-rich platforms and thoughtful aggregators online, but it expresses it’s intelligence specifically through music recommendations.
[Track 5: Everyday is Like Sunday by Morrissey]
I like to spar with the Spotify algorithm by listening to music I wouldn’t normally listen to and see how it generates playlists based on those very un-me choices. For example, I hate Morrissey a lot. But, if I start listening to him and the Cure, Spotify thinks I like them and will create a playlist for me called “Depressing Yowling for Sad British Teens” or something.
As unscientific as my approach is, it does have technical implications. This type of intentional “culture jamming” is known as “poisoning the dataset” and it affects other people’s experience, not just your own. In my case as a long time and very frequent Spotify user my choices can have a significant blast radius and is more effective the weirder it is. It could have 10–20x the impact of honest data because it’s such a statistical outlier.
My evaluation criteria are focused mostly on what I would describe as “vibe,” which is a soupy combination of BPM, instrumentation, and genre. I am admittedly biased against female vocalists.
[Track 6: Hallelujah by Jeff Buckley]
It goes like this…
- I decide who I would like the internet to think I am.
- Then, I choose three songs I think that person would like.
- Then, I add the first three songs that Spotify recommends to the playlist.
- Then, I add three songs I think are exactly like those three songs and so on until I get to 12 songs. Then, I listen. As I listen, I add songs in a similar fashion.
This playlist was one I created based on a version of myself I call “Ernie.” Ernie is a chill dude who works at a health food store and is into energy healing. He watches non-contact sports, camps and can often be found at his local vegan cafe drinking natural wines.
This playlist was one I created for a shadow of my current identity I call “Bronte.” They are a host at Dumbo House and really don’t care where your name is on the list. They have never been upstate, have very nice headphones and read a lot. They also volunteer on Sunday nights because they sleep til 3.
I do none of these things. But, somewhere on a server farm in midwest Idaho, there is a Spotify algorithm dreaming of a different me listening to music on a porch that doesn’t exist and a club I’d never get into. And I feel free.
If you create a playlist to experiment with hacking the algorithm please share it with me and I’ll listen to it all the time.
