Social media platforms are increasingly vulnerable to election interference campaigns. Within the context of the 2020 U.S elections, and increased domestic efforts to disseminate misleading political advertising, this Capstone project investigated policy and enforcement loopholes across large social media companies. The research consisted of a three-tiered approach: Platform Policy Comparisons, Ad-library Investigations, and Edge Case Ad Testing.

In the first approach, the Capstone team compared advertising policies across Google, Twitter, Facebook, Snapchat, and Reddit to understand each company’s definition and regulation of political and social issue ad content. The second approach investigated Google’s and Snapchat’s ad libraries to understand the scope of political advertisers, their spending and reach, and their advertisement content. In the third approach, the Capstone team created a left-leaning news opinion website populated with posts pertaining to climate change, immigration, abortion, and the 2020 election. Using thier opinion pieces, the Capstone team ran text and merchandise advertisements designed through a five-tiered system based on escalating political language. Starting with the least politically-charged, the team intensified the language after the prior ad passed review. This approach enabled the team to determine if unverified sources could publish political advertisements, the degree to which politically-charged language could be published before getting flagged, and whether merchandise and text ads are held to different levels of scrutiny.

Key observations include: policy carve outs for news and political merchandise are exploitable; misrepresentation went unflagged on both advertiser and content levels; election restrictions were inconsistently applied within their account and across ads; and user experience insufficiencies hinder advertisers from adhering to policies. Based on these findings, the Capstone team provided recommendations to tighten policy and enforcement to prevent election interference.