A corporation such as Spotify was founded a few years ago by a 23-year-old man, and it now plays a key role in the music industry. YouTube is now the big-boss of music delivery. Had I predicted these things ten years ago, you would have thought me odd.
Kurzweil, the inventor, popularized the notion of a technological singularity. It is a moment in our history where things move so fast, we can no longer comprehend the world nor make even short-term predictions. Are you game to predict the next ten years of the music industry?
Let us think about the present.
There is an explosion of diversity because it becomes affordable. Spotify can offer millions of songs. This was unthinkable in the past. The net result is that more musicians than ever can be found by more people than ever.
Chris Anderson in 2004 promised us a world where more of us could make a living on the “long tail” (the long stream of less popular products and services) without needing a big break. Is it what happened? I think we still debate it. Recently, a study on the Google Play store found that it is a super-star market where the bulk of the money ends up in the pocket of the few, with most people hardly making a penny. Should we blame the bulk of the players for their bad luck?
Having your data stored in the index is not enough to be found. Being visible is not the same thing as being found. When you use Google to search for something, it can return 20,000 relevant pages, but you are unlikely to look at more than 3 or 4. Tools like Spotify are not different. They use recommender systems or, in other words, AI, to help us find what we are looking for. They think for us.
People don’t always realize that tools such as Google know about you and provide personalized results. These tools are mirrors, necessarily imperfect ones. They are not neutral.
What do I mean by neutral? When using polls to predict election results, we measure what people think, but we also change what people think by publishing the results. A top-10 list in music is not just a mirror, it is also an active agent that changes what people listen to.
The rules of engagement also matter. Our democratic systems are often setup to favor the emergence of two dominant parties. It is the intended effect. It couldn’t be more obvious in the USA. Yet who is setting up the rules in the digital world? Do you know?
The big players like Google or Spotify have access to a remarkable wealth of data. They know who you are, where you are and what you do. They know your age, your race, your sexual orientation. No business could have dreamed of having so much data 30 years ago.
As such, it is not harmful.
But any complex and powerful system can have unexpected and undesirable effects. For example, “there’s a general tendency for automated decisions to favor those who belong to the statistically dominant groups.”
At a minimum, I think we should study what systems like Google and Spotify do so we can discuss it openly.
Problems don’t necessarily arise because the big players are malicious. Datta et al. (2015) found that online ads for highly paid jobs tended to be shown more often to men than women. It seems that the reason is that younger women are a prized demographic so they get ads from advertisers with deeper pockets instead.
What could we do in concrete terms? We could use volunteers who agree to have some of their Internet interactions monitored. We could also proceed with automated sousveillance, where we create fake accounts to keep an eye on the big players.
We already keep track of journalists, to check whether they are being fair, I think we should spend a bit more time tracking Google, Apple, Amazon, Spotify… This would seem only prudent in such a fast changing world.