Thrilled today to bring you another piece from a scholar I met at our WGS South conference in March, Hannah Steinhauer. Hannah is a Ph.D. student at Virginia Tech's Alliance for Social, Political, Ethical, and Cultural Thought. She’s from Portsmouth, VA and has lived in the South all her life. She has degrees in Math and Gender Studies, and in her free time likes to cook, read fantasy novels, and watch reality dating TV.
A 2017 study1 conducted by the Guttmacher Institute and UC Berkeley found that in a 32 day period in 2017, terms related to self-abortion (such as “at-home abortion,” “abortion pill,” and “how to miscarry”) were searched about 210,000 times in the United States. The researchers surveyed 1,235 of these individuals in order to learn more about what information they were hoping to find. 96% of people searching about self-abortion were female, and over 80% were 24 or younger. They were also disproportionately Black, Latinx, and from the South. 62% of respondents were pregnant and did not want to be, 10% were pregnant and unsure if they wanted to be, and 15% were Googling on behalf of someone else who was pregnant. Notably, many were not aware of the legality of abortion in their state.
These findings reveal three important things. First, that there was a significant lack of accessible information about abortion in the U.S. even when Roe still stood. Second, the number of people in the United States who use the internet to search for information about abortion greatly exceeds the number of people who actually obtain abortions. Finally, these results show that people use the internet frequently to find information about how to obtain abortions, both legal and illegal, self-induced and doctor-induced, whether the internet is a reliable source for this information or not.
In thinking through the limits and possibilities of finding reliable abortion information on the internet in the U.S. post-Roe, it is important to understand what we mean when we say the internet, and by extension, who owns this “internet.” When the internet first became widely available to the public in the 1990s due to its privatization2, there was a period of time in which essentially anyone could make a website or blog with very little dependance on large companies like Google or Microsoft. However, like many other commodities, large companies quickly gained a monopoly over internet technology. By the early 2010s, what we now know as the internet— search engines, social media platforms, etc.— is essentially owned by 6 companies, Google, Meta, Microsoft, Apple, Amazon, and IBM, and these companies build and govern the internet for most of the globe, excluding China. Feminist technology scholars have shown this to be an example of Patricia Hill Collins’ matrix of domination.
In the post Roe world, it appears that there may be consequences, both intended and otherwise, for those who attempt to search abortion access information online. In addition to pushes for increased police use of personal search histories to charge people with obtaining illegal abortions in states like Texas, there is an alarming trend of pro-abortion content being flagged or removed on Meta’s platforms Facebook and Instagram in the summer of 2021, leading up to and directly following the overturn of Roe. Users have seen posts about abortion removed with vague messages that the posts “violate community standards,” or have had sensitivity screens added to posts for abortion rights that label the posts as potentially “graphic or violent content,” even if the images in question are not graphic, as seen in the image below.
Facebook’s speech policies, which are found on Meta’s “Transparency Center” webpage, only mention abortion once, under the “Violent and Graphic Content” section. The policy reads, “For the following content, we include a label so that people are aware the content may be sensitive: Imagery of fetuses and newborn babies that show: Dismemberment. Visible innards. An abortion or abandonment context.” While this may have the good faith intention of preventing the common tactic of pro-life organizers showing graphic imagery of fetuses, it leaves open the question of what written content on abortion is allowed.
The mysterious removal and sensitivity screening of non-graphic abortion content reveals how precarious free speech is online when there are so few power holders in charge of internet companies, and how severe the consequences can be when there is a lack of legal accountability for why information posted by individual users can and is taken off of the site with little to no explanation. On June 28, 2022, the account @InstagramComms tweeted, “We're hearing that people around the world are seeing our "sensitivity screens" on many different types of content when they shouldn't be. We're looking into this bug and working on a fix now.” Algorithms are, simply put, a set of instructions for making sense of data, so an algorithm would not be capable of “deciding” what content is or is not in line with a speech policy outside of what it is trained to flag. I argue that the distinction between algorithmic and human content moderation3 is actually a false binary, because humans design algorithms and they cannot make decisions outside of the way that they are programmed. This means for this case study that either the content moderation algorithm(s) employed on Facebook and Instagram had to have “abortion” as a key word that would warrant a flag, or that humans made the decisions to censor abortion content following the overturn of Roe, both of which are the result of human decision making. Therefore, this binary allows Meta, and other tech companies, to use their content moderation algorithms as a scapegoat when people disagree with their decisions. If they can say that the algorithm had a “glitch” or a “bug” that resulted in a decision people find inappropriate or an act of censorship, this removes individual accountability.
Censorship is not inevitable, and should be seen as something humans do, not computers. It is not a coincidence that one of the few white men that owns much of our internet technology, Elon Musk, is a pronatalist who recently donated $10 million to a fertility research institution in Texas. However, all is not lost. Many abortion activists, specifically those located in states with the greatest abortion restrictions, are already viewing abortion activism on the internet as not guaranteed moving forward. Many, including in South Carolina, are turning to second wave feminist strategies like hotlines and zines, that predate the internet. Historically, feminists have found unlikely strategies to distribute necessary reproductive information and services to those who need them. It is humans who censor abortion information, but it is also humans who create and maintain reproductive justice.
Jenna Jerman, Tsuyoshi Onda, and Rachel K. Jones, “What are People Looking for When They Google ‘Self-Abortion’?” Contraception 97, no.6 (June 2018): 510-14.
J. Abbate, "Privatizing the Internet: Competing Visions and Chaotic Events, 1987–1995," in IEEE Annals of the History of Computing, vol. 32, no. 1, pp. 10-22, Jan.-March 2010, doi: 10.1109/MAHC.2010.24.
Roberts, Sarah T. Behind the Screen: Content Moderation in the Shadows of Social Media. Yale University Press, 2019.
Certainly, a lot of ROWING AND WADING...but we never give up!
Excellent story and nicely aligned with Julia Anguin's post about Google's use of AI in search: https://www.proofnews.org/lessons-from-googles-ai-mistakes/