UP NEXT: YouTube's Recommendation System and the 2019 Canadian Federal Election
In the months leading up to the 2016 election in the United States, YouTube's recommendation algorithm decidedly favored pro-Trump videos, fake news and conspiracy theories. In this thesis, I question whether such bias is present in the context of the 2019 federal election in Canada. To do so, I make use of open-source software to gather recommendation data related to three of the candidates: Conservative Party of Canada leader Andrew Scheer, New Democratic Party leader Jagmeet Singh, and
... et Singh, and Liberal Party of Canada leader Justin Trudeau. Using the same data, I will also study the media bias and factual accuracy of the sources recommended. My results show that YouTube's recommender system is susceptible to influence by audiences and shows bias towards Andrew Scheer and against Justin Trudeau. Given my results and evidence provided by other researchers, this study stresses the need for ethical algorithm design, including proactive approaches for increased transparency, regulatory oversight, and increased public awareness. iii ACKNOWLEDGEMENTS I would like to extend a big thank you to my supervisory committee, Tami, Astrid, and Harvey for keeping me inspired and asking the right questions. Your countless hours of work have not gone unnoticed, and through your efforts you've made me a better researcher and writer.