LGBTQ Content Filtering: Final Summary Post

This project has become something much different than what I originally set out to do. Although I wanted to focus more on the algorithmic aspects of these issues, I underestimated how secretive Facebook is with specifics like these. I still have learned a great deal about the issues still facing the LGBTQ community on Facebook, and what still needs to be done. Although it seems like the company truly does have the interest of the community in its heart, no policy or solution is perfect. That’s why it is still so important to keep researching and fighting and trying to make this platform work for everyone. It is important to remember that these issues are not just about quality of life; problems like privacy and data protection can be life and death situations, especially in places where queer people are oppressed. Facebook has the power to help the LGBTQ community be safe and thrive online, and we just have to push them to keep moving forward.

I have looked at LGBTQ ads in the Facebook Ad Library. Out of 31 different advertisers looked at, 78 of total 167 ads had been at some point removed for not having the “social issues, elections, or politics” disclaimer. Out of the total of $13,421 dollars for these ads, about $7,498 of that was spent on the ads taken down for no disclaimer. Although Facebook does not provide information on how long these ads were supposed to run for, this is money spent by advertisers that, because of the penalty for not having the disclaimer, was not fully utilized. Many of these ads had nothing political about them; they were for events, therapy sessions, history exhibits, etc. The way Facebook automates the ad process affects the most vulnerable in the LGBTQ community, keeping us from using the platform as equals.

I really think Facebook should remove “social issues” from the “social issues, elections, or politics” category entirely. The machine learning classifier used when checking advertisements needs to be able to distinguish between political and non-political ads. If an advertisement makes no mention of a politician, current legislation, or voting in an election, it poses much less of a threat. 

I also looked at Micro-Targeting, and how the LGBTQ community can be targeted through Facebook’s platform. Between fake ads ran by the Russian government, to conversion therapy ads being targeted at gay men, to the threat of Facebook’s data being forced to be shared, the community is not safe from danger on social media. There needs to be more oversight when targeting certain demographic groups. The platform should be able to recognize that allowing content like conversion therapy to be advertised is dangerous to the LGBTQ community, and should not be allowed to be ran. Although it can be beneficial for advertisers to be able to more easily connect with the queer community, information on users’ gender identity and sexuality needs to be respected and protected.

This project has been very fun to do, and has definitely piqued my interest in possible future research areas. If anyone is interested, here is a link to my final paper: Monroe 2019 Research Paper. Thanks for reading!