Discussion

How does ML affect pop culture?

ML affects many aspects of pop culture, including as a recurring theme in movies and books, a driving force behind online sales, and part of the production of culture by popularizing trends, or the actual creation of content. The three facets of pop culture that were examined in this exhibit are impacted by ML in ways that have transformed the cultural industries. Artists can use ML to generate new ideas, pinpoint a target audience or predict upcoming industry trends. ML is used to predict what movies, books and songs will be popular, which by extension predicts where producers and publishers invest. The cultural labour forces are also impacted as certain jobs are taken over by ML, or modified to rely heavily upon ML. 

ML also affects how pop culture is consumed. Cable television is increasingly being replaced by video streaming services, AM/FM radios are swapped out for audio streaming, and bookstores are closing as more consumers rely on Amazon for their purchases. Academically, ML is being used to consume and analyze culture. Computers are “watching”, “reading”, and “listening” to huge catalogues of cultural artifacts, and with this data ML is able to provide the basis for critical analysis. Having computers perform time consuming tasks, like watching a corpus of movies, saves scholars valuable time, thus allowing research to advance at an increasingly rapid rate. However, the biggest impact of ML on the cinematic, literary and music industries is the use of recommender systems.

Recommender systems are big business and the key to success to many digital platforms. Recommenders take large swaths of user generated data to narrow down choices on a platform, and present personalized and curated content. As explained in the three case studies, some services offer better recommendations than others, which ultimately affects user uptake and retention. For example, Netflix has a very high client retention rate (Gomez-Uribe & Hunt, 2016) which is attributed to a well programmed recommender system, despite being a paid service. The way a ML recommender is programmed will affect what is deemed “popular” on a platform. Popular content will sell more products, and attract new users. This will influence future content, as producers and publishers look to past successes to create future successes. Algorithms can also be programmed to promote specialized products that may be of interest to the user, to increase a business’ bottom line, or likely both.

Further research

While out of the scope of this project, there were elements of ML and pop culture that were not explored in the exhibit but warrant further research. One such example is the larger scale impact that the use of AI could have on the human labour force of cultural industries. The exhibit has shown how ML is being used to make movies, create music, and write books, but currently these tools contribute to and improve human work, rather than replacing human work. Another area that this exhibit did not discuss in depth was the effect of ML algorithm bias on marginalized populations. The case studies highlight examples of ML bias and misrepresentation in recommender systems’ results, however bias is common among ML in pop culture (and in general), which can be damaging and harmful.

The world of ML is vast, and this exhibit has only skimmed the surface. The three areas of pop culture that were discussed are not the only industries that are affecting by AI and ML. As ML programs being increasingly part of our daily lives, it is important to understand the technology that we rely upon to get us through our days. We hope you have learned something new about ML in order to spot it, question it, and ultimately improve it.