by: Anna Transou
Within the recent decades, humankind has experienced a massive shift from how we expect media content to be presented to us. Alongside the introduction of artificial intelligence and supercomputers, the modern human seems to be quite comfortable with its current relationship with the algorithms that act as our own personal shopping or movie recommendation assistants. But how much do we really know about exactly how we’re being prescribed content, and why haven’t we all taken a step back to even think about the further implications of a computer generated culture such as the one we’re headed to, if not already in now? To answer some of these questions I have homed in on Netflix, the film and television platform with a whopping 213.6 million subscribers, to dive into the implications of humanity complacently entering into an agreement with computerized technologies in the supposed “co-creation” of culture.
Despite Netflix users interacting with and benefitting from the personalized suggestions that the Netflix Recommendation System (NRS) offers to us, what do we really know about how it works? Regarded as “trade secrets”, the complex algorithms are technically incomprehensible by the average consumer, but we do know some things about how they are designed and intended to function. In a 2014 article containing an interview with Todd Yellin, VP of Netflix product innovation, The Atlantic discovered more about how the media platform has accrued a “stockpile of data about Hollywood entertainment that is absolutely unprecedented.” In a process to reverse-engineer Hollywood, Netflix used large teams of people to deconstruct a myriad of elements within a vast collection of film and television to create microtags which would then help sort content into roughly 76,897 microgenres. Yellin boasted to The Atlantic about the efficacy of the algorithm’s personalization which aims at “putting the right title in front of the right person at the right time,” illustrating the formula’s analogous nature to that of Facebook’s highly customized NewsFeed. In conjunction with the accumulation and implementation of user data in creating personalized suggestions, the creation of NRS involved a process of probing the content individually, inspiring an investigation for the implications for algorithmically curated content.
The overarching concerns lie in the vexing unknown of what it means to involve computerized algorithms in the creation, definition, and curation of culture, a domain that has historically been reserved for humans. Netflix and many other algorithmically controlled platforms are primarily withholding a majority of the power in the distribution of cultural content, despite what their parent companies advertise to users. In the specific case of Netflix (but also applicable to many other media platforms), “reality gets coded into data for the machines, and then decoded back into descriptions for humans.” Such concerns over who has the power and authority to shape culture and ideas dates back at least to Marx and Engels’ discussion of the ruling class’ control of media and mental production. Whereas the curation of culture was once solely determined by the upper classes of whom held the means to produce media, today we are placing the significant task of cultural prescription into the inanimate hands of artificial intelligence. The construction and suggestion of cultural tastes by algorithms repositions human autonomy as, potentially, a secondary force, in which Netflix users provide the system with ample consumer data to facilitate the operation of recommendations in a one-way distribution of media content. The imbalance of power between human and machine occurs in the restraints placed upon users in that there is no true recourse for the consumer to edit their viewing history, interact socially with others on the platform, as well as the limited binary rating system which merely asks for a user to give content a thumbs up or down. Communication scholar, Fatima Gaw illustrates this dynamic further in pointing to the “restrictions [that] exist in the same paradigm where users are supposedly empowered by free choice and self-determination of their cultural experiences.” After all, Netflix promotes that “our members control what they watch” despite the constraints implemented within the platform to limit user agency in constructing their own individualized taste preferences apart from the algorithm’s prescriptions. The rigidity of the NRS “restrict[s] and often supersede[s] user agency” as it operates neither in an emancipatory or deterministic manner, but instead “inhibits users from deviating from the formula it prescribes.”
With proprietary goals in mind, Netflix’s recommendation system was not built to encourage diversity of media within the platform, instead, the algorithm operates to feed content to users similar to what they have already been favorable towards. The system’s primary functions bring up concerns of the development of cultural echo-chambers, such as has been discovered on social media apps like Facebook and Instagram. Further research must be done in order to discern what effect media echo-chambers will have on the development of culture and society, and how humans can fight back against an algorithmically produced mono-culture of the future that lacks cultural diversity.
