What happened when I let algorithms run my life for a week
Algorithms are everywhere. They rule our lives. They decide what we see on social media, which ads get targeted to us, which route we should take to get places. If it used to be The Man keeping us down, now it is The Algorithm (or, perhaps more accurately, The Man’s Algorithm, but you get the idea).
For one week, I decided I would try to take a more conscious look at the kind of mundane algorithms I use everyday. From scrolling social media when I first wake up in the morning to route-planning my way home in the evening, I wanted to make myself more aware of the algorithms that have become routine in my daily life, how they affect my choices, and what they reveal about me.
I would attempt to live my life “by algorithm” – using algorithm in the broad, colloquial sense here to mean any set of computer calculations that results in a solution, including recommendation engines, filtering systems, prioritisation or personalisation algorithms and so on. I would track my interactions with algorithms and let them make my choices for me.
If Google Maps said I should take a certain route, I would. If Netflix suggested I watch a certain film, I would. Basically, when I came to a decision that an algorithm could make for me, I would turn to it – like the book Yes Man, except instead of saying yes to everything I’d just let the algorithm answer on my behalf.
READ NEXT
Machine learning is totally changing what we think of as literature
Machine learning is totally changing what we think of as literature
By WILL BEDINGFIELD
Armed with smartphone and laptop, plus a vague hope of outsourcing some of the everyday human burden of decision-making to technology, I set off.
Day 1
On the first day of my experiment, I immediately realise a downside of diarising my algorithmic interactions: this is going to reveal more about me than I had perhaps appreciated. Based on my usual routine, I’d imagined the first algorithm I’d come across in the week would be on Instagram or Spotify, but it’s not. It’s the navigation system in my dad’s car as he takes me to Topps Tiles before work to choose some tiles for my bathroom, which he’s helping me fix up.
The navigation system maps our route; it’s one of those that has a woman’s voice narrate every upcoming turn. We get the tiles, and I head to the station to go to work.
I always take exactly the same route to work, which is the one Citymapper told me I should take when I first moved to the flat I’m currently living in. In the spirit of my experiment, I check the app anew to see what it suggests. My usual route, a mix of Overground and Underground, comes in as fastest, so I stick to it.
On the way, I open Spotify. I’m particularly interested to see how recommender systems judge my tastes, so instead of picking one of my own playlists I scroll down to the “Made for…” section on the home screen and select the first playlist supposedly designed for me. It’s “Release Radar”, which claims to include new releases “picked just for you”.
READ NEXT
The natural genius of ants is helping us build better algorithms
The natural genius of ants is helping us build better algorithms
By MATT REYNOLDS
The first song really reflects much better on my musical taste that I could reasonably have hoped for (I listen to a lot of terrible music); it’s a remix of a Hot Chip song. Next up is a new single by Madonna. Maybe I should let Spotify pick my music more often.
The third song, however, reveals a fatal flaw in the system: it’s a classical track. I don’t listen to classical much, but I do have a classical playlist I sometimes go to for background music when I’m doing quiet work, which is presumably why this song has found itself recommended to me. It’s a dramatic arrangement featuring choir and strings, but it’s not the right vibe.
This is possibly the biggest issue with my Spotify recommendations: I listen to different types of music at different times. My most listened to playlists are my gym playlist (cheesy motivational tracks) and one that I call “guilty pleasures” (nostalgic 90s and 00s pop and indie). But while I may listen to both of these embarrassingly often, I don’t like to think that they represent my “real” music taste, and I don’t particularly want to discover other songs or artists similar to them.
As if to prove my point, the next new release I’m recommended is by Blink-182.
For one week, I decided I would try to take a more conscious look at the kind of mundane algorithms I use everyday. From scrolling social media when I first wake up in the morning to route-planning my way home in the evening, I wanted to make myself more aware of the algorithms that have become routine in my daily life, how they affect my choices, and what they reveal about me.
I would attempt to live my life “by algorithm” – using algorithm in the broad, colloquial sense here to mean any set of computer calculations that results in a solution, including recommendation engines, filtering systems, prioritisation or personalisation algorithms and so on. I would track my interactions with algorithms and let them make my choices for me.
If Google Maps said I should take a certain route, I would. If Netflix suggested I watch a certain film, I would. Basically, when I came to a decision that an algorithm could make for me, I would turn to it – like the book Yes Man, except instead of saying yes to everything I’d just let the algorithm answer on my behalf.
READ NEXT
Machine learning is totally changing what we think of as literature
Machine learning is totally changing what we think of as literature
By WILL BEDINGFIELD
Armed with smartphone and laptop, plus a vague hope of outsourcing some of the everyday human burden of decision-making to technology, I set off.
Day 1
On the first day of my experiment, I immediately realise a downside of diarising my algorithmic interactions: this is going to reveal more about me than I had perhaps appreciated. Based on my usual routine, I’d imagined the first algorithm I’d come across in the week would be on Instagram or Spotify, but it’s not. It’s the navigation system in my dad’s car as he takes me to Topps Tiles before work to choose some tiles for my bathroom, which he’s helping me fix up.
The navigation system maps our route; it’s one of those that has a woman’s voice narrate every upcoming turn. We get the tiles, and I head to the station to go to work.
I always take exactly the same route to work, which is the one Citymapper told me I should take when I first moved to the flat I’m currently living in. In the spirit of my experiment, I check the app anew to see what it suggests. My usual route, a mix of Overground and Underground, comes in as fastest, so I stick to it.
On the way, I open Spotify. I’m particularly interested to see how recommender systems judge my tastes, so instead of picking one of my own playlists I scroll down to the “Made for…” section on the home screen and select the first playlist supposedly designed for me. It’s “Release Radar”, which claims to include new releases “picked just for you”.
READ NEXT
The natural genius of ants is helping us build better algorithms
The natural genius of ants is helping us build better algorithms
By MATT REYNOLDS
The first song really reflects much better on my musical taste that I could reasonably have hoped for (I listen to a lot of terrible music); it’s a remix of a Hot Chip song. Next up is a new single by Madonna. Maybe I should let Spotify pick my music more often.
The third song, however, reveals a fatal flaw in the system: it’s a classical track. I don’t listen to classical much, but I do have a classical playlist I sometimes go to for background music when I’m doing quiet work, which is presumably why this song has found itself recommended to me. It’s a dramatic arrangement featuring choir and strings, but it’s not the right vibe.
This is possibly the biggest issue with my Spotify recommendations: I listen to different types of music at different times. My most listened to playlists are my gym playlist (cheesy motivational tracks) and one that I call “guilty pleasures” (nostalgic 90s and 00s pop and indie). But while I may listen to both of these embarrassingly often, I don’t like to think that they represent my “real” music taste, and I don’t particularly want to discover other songs or artists similar to them.
As if to prove my point, the next new release I’m recommended is by Blink-182.
Recommended News for You
The key to creating virtual conferences that work
15 February 2023
Netflix’s Free Weekend, StreamFest
20 November 2020
Art Paris spring fair finally goes ahead
20 November 2020
Google is finally fixing one of Chrome's biggest privacy flaws
07 June 2019John
