Microsoft Word - Monea-The Digital Closet-The Digital Closet.docx


 
Selected Papers of #AoIR2021:  

The 22nd Annual Conference of the  
Association of Internet Researchers 

Virtual Event / 13-16 Oct 2021  
 

 

Suggested Citation (APA): Monea, A. (2021, October). The Digital Closet: Platform Censorship and 
LGBTQIA+ Independence Online. Paper presented at AoIR 2021: The 22nd Annual Conference of the 
Association of Internet Researchers. Virtual Event: AoIR. Retrieved from http://spir.aoir.org. 

THE DIGITAL CLOSET: PLATFORM CENSORSHIP AND LGBTQIA+ 
INDEPENDENCE ONLINE 
 
Alexander Monea 
George Mason University 
 
 
This paper draws on research for my forthcoming book – The Digital Closet: How the 
Internet Became Straight from MIT Press – to examine the coders, data, algorithms, 
and policies that govern the circulation of content on internet and technology platforms 
like Google, Facebook, Apple, and Amazon. I argue that these platforms are 
increasingly embedding heteronormative and cisnormative bias into their algorithms, 
datasets, and policies that are censoring LGBTQIA+ speech across the internet. This 
presentation combines archival studies, discourse analysis, platform studies, and critical 
code studies to piece together an analysis of what is going on inside the proprietary 
software at Google and Facebook in particular. 
 
The first section of the paper draws on scholarship on Silicon Valley culture and ‘toxic 
geek masculinity’ to analyze the endemic heteronormativity and cisnormativity at work in 
tech culture. I connect these critical-analytical scholars to data from the Department of 
Labor produced during a protracted legal battle of sexist hiring and retention practices at 
Google, testimonials from former Google employees, and informal surveys and screen 
captured conversations in the wake of James Damore’s infamous ‘Google memo’ circa 
2017. By reading across these various sources I develop an understanding of Google’s 
work culture as having deeply embedded normative understandings of gender and 
sexuality that are directly connected to the company’s repeated failure to anticipate and 
meet the needs of LGBTQIA+ users. 
 
The second section of the paper draws on original investigations into the benchmark 
datasets that Google used to develop its image recognition algorithms (like the GoogLe 
Net algorithm behind Google Photos and SafeSearch). I provide examples from 
WordNet and ImageNet to demonstrate the substantial heteronormative and 
cisnormative biases that are embedded in these datasets and provide some examples 
of the ways in which these biases are amplified once they are learned and applied by 
machine learning algorithms like the convolutional neural networks used for computer 
vision and image recognition tasks. 
 



 
The third section of the paper draws on original investigations into Facebook’s content 
moderation policies by performing close readings of its Community Standards 
Enforcement Reports, documents from its Product Policy Forum (previously called the 
Content Standards Forum), and leaked documents obtained by investigative journalists. 
Throughout this analysis I demonstrate the systemic heteronormative and cisnormative 
bias that structures Facebook’s community standards and content moderation efforts on 
the platform and connect them directly to its attempts to produce “human algorithms” 
through outsourced digital piecemeal labor. 
 
In closing, I argue that while less definitive than desirable due to the proprietary nature 
of platform code, these analyses demonstrate a substantial heteronormative and 
cisnormative bias across some of the most popular internet platforms in the world. As 
they increasingly dominate web traffic, they begin to structure our very experience of the 
internet, inflecting it in a heteronormative and cisnormative way. In so doing, they are 
damaging the capacities the internet might otherwise have for LGBTQIA+ independence 
and the fostering of community building, social organizing, sex education, and the 
exploration of non-normative erotics and desires online.