Taste for Algorithm

The platform economy is not just about brewing data. It transports them, directs them and redirects them into the jungle of online content thanks to these recommendations called algorithms. Faced with the overabundance of flows, the evidence seems to impose itself: the need to identify oneself creates the system that facilitates choice. Yet, far from being limited to the accompaniment of a decision, the reign of algorithms is not without consequences on cultural diversity and the construction of tastes. Algorithms orientate the people with online products to adjust them depending on all the data poured into the platforms. As they seem to know us better than we know ourselves, they direct us without asking us in the infinite amount of content. These are the perfect extensions of our ego, our image getting out there.

One figure, again one, should inspire reflection. He also comes from the United States, where 99% of the listening is mobilized on 20% of the catalog of the platform Spotify. He especially denies the prophecy of the digital “long tail”, theorized in 2004 by Chris Anderson, who wants products that are not popular, in this case cultural, to wait for their audience. “A cultural work is a good experience, whose quality can not be appreciated without approaching it,” says economist Pierre-Jean Benghozi, a specialist in digital and cultural and creative industries. “Symbolic value is as important as usage value, and you never know where and when success will come. » What about the platforms? Specialists point a paradoxical double movement in which the contents are fragmented and the markets are fragmenting, while generating concentration and the star system. “The Internet has a more structural than quantitative effect,” continues Pierre-Jean Benghozi. The channels are multiplying, through downloading and streaming. In this context, the works are fragmented and this is particularly true in music. We listen to songs and not albums. From then on, the offer is reorganized and the scale changes. The consequences of this restructuring of the offer via the Internet upset an entire sector.

Logically, publishing, or rather the “editorialization” of content by platforms, weakens traditional record majors or conventional producers. A myriad of short circuits develop where authors and artists self-publish immediately. In an increasingly fragmented market, there is the dual effect of the proliferation of offers for identical content, and at the end, what Pierre-Jean Benghozi calls the passage “from abundance to hyper-supply”. It is precisely here that algorithms intervene, without promoting the expected balance in perpetual surplus flows. Literally, algorithm is corresponding to a sequence of instructions. A mathematical instruction given upstream produces a suggestion of result downstream. In the digital world, it is from the strategic weight of the data collected that the algorithm becomes a vector of value creation. Content a priori plebiscite, and therefore profitable according to the law of the platform economy, will be all the more likely to be passed on. So we observe that less than 1% of the content is delivered in the majority of the cases. The instrument is one thing, its handling is another. Support ultimately has less than the latitude that is allowed. Are we afraid that artificial intelligence will soon come to go beyond human intelligence?

The algorithm is not that benevolent, volatile and floating adviser to painless recommendations. Programmed and intentioned, it provokes the conjunction, even the coincidence, between the quest of the user sowing his data and the interest of the prescriber — the platform — dissecting them. The process might be harmless after all as “listeners do not know what they are looking for”. The algorithm would surf at the same time on the anticipated desires and the part of indecision of the users. By encouraging the listener to listen to the highest bidder, he would entertain him in the illusion of satisfying his tastes.

What would remain of culture without diversity and, ultimately, with no other creation than the reproduction of models encouraged by the digital industry? We have to understand technology to expand our views — not about caging ourselves.