DS Extra Readings & Case Studies

(Mayworm, Albert, Haimson, 2024)

• transgender bodies make visible cisnormative content moderation policies and enforcement in a meta oversight board case

• Transgender and nonbinary social media users experience disproportionate content removals on social media platforms, even when content does not violate platforms’ guidelines.

• In 2022, the Oversight Board, which oversees Meta platforms’ content moderation decisions, invited public feedback on Instagram’s removal of two trans users’ posts featuring their bare chests, introducing a unique opportunity to hear trans users’ feedback on how nudity and sexual activity policies impacted them.

• Commenters criticized Meta’s nudity policies as enforcing a cisnormative view of gender while making it unclear how images of trans users’ bodies are moderated, enabling the disproportionate removal of trans content and limiting trans users’ ability to use Meta’s platforms.

•  the Oversight Board’s public comment process demonstrates the value of incorporating trans people’s feedback while developing policies related to gender and nudity, while arguing that Meta must go beyond only revising policy language by reevaluating how cisnormative values are encoded in all aspects of its content moderation systems

• META removed posts fundraising for top surgery

• Though most social media platforms forbid many types of nude or sexual content, many platforms exempt certain categories of nude content from removal, such as nudity in healthcare imagery.

• Cisnormativity is defined as “the assumption that all people are cisgender, i.e. [having] a gender identity and presentation that are consistent with the sex they were assigned at birth” [14], while transgender people and experiences are either unaccounted for or are considered “abnormal”

• social media corporations like Meta must seek feedback from trans policy experts while revising gender and nudity policies that disproportionately impact trans users; ideally, they should pre emptively work with trans policy experts while initially developing platforms’ policies relating to gender.

• Meta also allows exceptions to the “female nipple” ban in certain medical contexts (including gender-confirming surgery contexts) [19]. Meta refers to “female breasts” and “female nipples” to describe specific kinds of chest nudity, but the policy does not explicitly define these terms.

• no specific guid ance for moderating trans nudity, the posts were likely removed for containing “female” chest nudity– despite neither user being female, and neither post showing exposed nipples. GREY AREA

• Trans people uniquely rely on social media to meet their trans specific needs, such as seeking out trans healthcare information [3], crowdfunding for gender-affirming healthcare [4, 21], visibility and activism [12], expressing trans identity [16], and finding community amongother trans people

• Further, policy language that uses binary language to describe body parts does nothing to inform nonbinary users how content featuring their bodies will be moderated

(Mayworm, Albert, Haimson, 2024) - Burdens and change

• Overall, the commenters agreed that Meta’s disproportionate moderation of trans users and their content (enabled by Meta’s cisnormative nudity policy and ineffective algorithmic detec tion of gender) results in trans users and their bodies being erased on Meta’s platforms, preventing trans people from using Meta’s platforms as freely as other users or to meet their trans-specific needs.

• Instead of focusing on Meta’s differing gendered moderation of nudity, these commenters argued that removing gendered descrip tions of body parts from Meta’s nudity policy may help prevent future incorrect removals of trans users’ content, while affirming trans users’ genders instead of describing their bodies in inaccurate and harmful ways.

• Overall, commenters argued that Meta should replace gendered descriptions of body parts with gender-neutral language in their Adult Nudity and Sexual Activity policy, allowing the policy to be more inclusive of trans users while preventing future incorrect removals of content featuring trans bodies.

• We argue that to avoid trans exclusion and erasure online, companies like Meta must go beyond policy language updates to also critically as sess and adjust how embedded cisnormative values inform content moderation and policy enforcement systems.

(Kojah et al)

Dialling it Black: Shadowbanning, Invisible digital labour, and how marginalised content creators attempt to mitigate the impacts of opaque platfrom governance

• Content creators with marginalized identities are disproportionately affected by shadowbanning on social media platforms, which impacts their economic prospects online.

• We highlight the labor and economic inequalities of shadowbanning, and the resulting invisible online labor that marginalized creators often must perform.

• 3 types of invisible labour: mental and emotional labor, misdirected labor, and community labor.

• even though marginalized content creators engaged in cross-platform collaborative labor and personal mental/emotional labor to mitigate the impacts of shadowbanning, it was insufficient to prevent uncertainty and economic precarity created by algorithmic opacity and ambiguity.

  Because it is difficult to detect or confirm, shadowbanning is less risky for platforms than removing content, yet it still enables platforms to maintain substantial power over users and content

• e.g. witter, which prevents shadowbanned accounts’ usernames from appearing in search [67]; TikTok, which has been accused of hiding Black creators from its For You page [50]; and, notoriously, by Instagram, which has had to apologize to swathes of users for restricting the reach and views of their content [30]. A key characteristic of shadowbanning is that, often, platforms do not notify users it is happening (ibid)- a benefit and a hindrance.

• shadowbanning as we know it may change or decrease in some parts of the world after the approval of the European Union’s Digital Services Act (DSA), which is the first major legislation to regulate the transparency of visibility remedies [55], making shadowbanning a legal matter.

• Creator economy is part of the gig and platform economy

• This opaque, inscrutable “algorithmic boss” is often viewed as the main entity responsible for the precarity of creator labor, an entity that can wreak havoc on creators’ lives and livelihoods [24]. This job insecurity triggers financial anxiety, which negatively impacts well-being [23, 29] in a field that already provides inadequate healthcare and lacks most forms of organizational support [78], creating a notable power differential between platforms and their workers [2] [30].

• hough creating digital content is not often viewed as work, it actually generates substantial financial benefit for platforms [49]- a form of online “free labor”

• To this end, social media users often find themselves developing and sharing ’folk theories’ about how algorithms work, so they can plan their work around them [28], a behavior that Bishop called “algorithmic gossip,” or a set of “communally and socially informed theories and strategies pertaining to recommender algorithms” [6]. Algorithmic systems are unstable due to the dynamic nature with which social media platforms change their values/morality; this shifting moral space means that users increasingly have trouble gaining control over systems that render them visible and that are tied to their identities

• Folk theories and algorithmic gossip are a known form of protection and defense, particularly for women and marginalized communities [6] [22]. However, they are also a form of knowledge generation which is not recognised, and sometimes even fought or undermined, by platforms themselves

EE - Shadow banning

One of the most notorious examples of shadowbanning was on Instagram, which previously had to apologize to pole dancers [30], admitting that a form of light suppression of hashtags and content was taking place [56], although Instagram’s CEO subsequently stated that ’shadowbanning is not a thing’ [16]. The platform then admitted that they do limit the visibility of specific content, such as content by Black creators and content showing people’s skin and bodies [30]. In recent times Instagram’s notorious censorship and de-prioritising of content depicting bodies has come under the scrutiny of Meta’s own Oversight Board, which overturned the platform’s ban of posts depicting transgender and non-binary nudity, criticizing its often confusing and discriminatory moderation of bodies and calling for clearer, fairer nudity governance that is consistent with international human rights and self-expression

• The content creators who participated in this study perceived that people with marginalized identities (e.g., women, pole dancers, and people whose bodies do not fit mainstream beauty standards) were more affected by shadowbanning and other forms of censorship than men who created similar content, and than people who fit conventional beauty standards.

• The gender disparities in content moderation were particularly acute across various platforms, with women pole dancers in particular perceiving that men who posted pole dancing content were not as heavily shadow-banned or censored.

EE - Queer shadow banning

• Participants suspected that the word “lesbian” was made less visible by platforms: for example, they described how when trying to search for lesbian content in TikTok’s search bar, the platform did not show hashtags like it did for other words, leading them to believe ’lesbian’ was shadowbanned. This would cause them to type it out with different spellings, i.e. “Le$bian”, which creates additional labor in their everyday platform use. We further explore Algospeak, and its connection to the invisible labor that content creators must perform, in a later Results section

• In these instances, content moderation seems to be blending with offline inequalities, replicating the status quo in the creator and gig economy rather than subverting it

• The ambiguity surrounding the reasons for shadowbanning directly correlates with when and how the content creators initiated shadowbanning

(Boyd and Crawford, 2012)

• Big Data is a cultural, technological, and scholarly phenomenon that rests on the interplay of technology, analysis, and mythology that provokes extensive utopian and dystopian rhetoric

• We have a new data ecosystem

(1) Technology: maximizing computation power and algorithmic accuracy to gather, analyze, link, and compare large data sets.

(2) Analysis: drawing on large data sets to identify patterns in order to make economic, social, technical, and legal claims.

(3) Mythology: the widespread belief that large data sets offer a higher form of intelligence and knowledge that can generate insights that were previously impossible, with the aura of truth, objectivity, and accuracy.

• Viewed by some as a powerful tool to address various societal ills OR  a troubling manifestation of Big Brother

• Necessary to ask which systems are driving these practices and which are regulating them

• The market sees data as an opportunity

• Big Data changes the definition of knowledge: changing the objects of knowledge, while also having the power to inform how we understand human networks and community

Radical shift about how we think about research: capacity to organise, collect and analyse data with greater breadth and depth

• Claims to objectivity and accuracy are misleading

Working with Big Data is still subjective, design decisions that determine what will be measured also stem from interpretation, Large data sets from Internet sources are often unreliable, prone to outages and losses, and these errors and gaps are magnified when multiple data sets are used together, To make statistical claims about a data set, we need to know where data is coming from; it is simi larly important to know and account for the weaknesses in that data.

• Too often, Big Data enables the practice of apophenia: seeing patterns where none actually exist, simply because enormous quantities of data can offer connec tions that radiate in all directions.

• Big data does not always mean better data: just because Big Data presents us with large quantities of data does not mean that methodological issues are no longer relevant

• Taken out of context, Big Data loses its meaning

Large data sets can be modelled, data is often reduced to what can fit into a mathematical model.

The rise of social netwrok sites prompted an indsutry-driven obsession with the 'social graph'

(Boyd and Crawford, 2012) - Their case study

• Twitter:

Twitter does not represent ‘all people’, and it is an error to assume ‘people’ and ‘Twitter users’ are synonymous: they are a very particular sub-set. Neither is the population using Twitter representative of the global population. Nor can we assume that accounts and users are equivalent. 

Furthermore, the notion of an ‘active’ account is problematic. While some users post content frequently through Twitter, others participate as ‘listeners’ (Craw ford 2009, p. 532). Twitter Inc. has revealed that 40 percent of active users sign in just to listen (Twitter 2011). The very meanings of ‘user’ and ‘participation’ and ‘active’ need to be critically examined.

It is also hard to understand the sample when the source is uncertain. Twitter Inc. makes a fraction of its material available to the public through its APIs.

• The ‘firehose’ theoretically contains all public tweets ever posted and explicitly excludes any tweet that a user chose to make private or ‘protected’. Yet, some publicly accessible tweets are also missing from the firehose. Although a handful of companies have access to the firehose, very few researchers have this level of access. Most either have access to a ‘gardenhose’ (roughly 10 percent of public tweets), a ‘spritzer’ (roughly one percent of public tweets), or have used ‘white-listed’ accounts where they could use the APIs to get access to different subsets of content from the public stream.3 It is not clear what tweets are included in these different data streams or sampling them represents. It could be that the API pulls a random sample of tweets or that it pulls the first few thou sand tweets per hour or that it only pulls tweets from a particular segment of the network graph.

(Boyd and crawford ) - Effects

• New networks

'articulated networks': result from people specifying their contacts through technical mechanisms

'behavioural networks': derived from communication patterns, cell coordinates and social media interactions

• Just because it is accessible does not mean it is ethical:

What is the status of so-called ‘public’ data on social media sites? Can it simply be used, without requesting permission? What constitutes best ethical practice for researchers? Privacy campaigners already see this as a key battleground where better privacy protections are needed

Accountability is needed and as part of a multi-directional relationship

• Limited access to big data creates new digital divides: data can be restricted within social media companies, some sell access for a fee, and others offer small data sets to university-based researchers

Unevenness because those with money can produce different research

Gendered and Race/Regional divisions: class of the big data rucg s reinforced through the univeristy system, those from the periphery are less likely to get those invitations

Complex questions about what kinds of research skills are valued and how these are taught

Whenever inequalities are explictly written into the system, they produce class-based structures

(Senft and Baym, 2015)

• The selfie as a cultural artifact and social practice. A way of speaking and an object to which actors (both human and nonhuman) respond.

• a selfie is a photographic object that initiates the transmission of human feeling in the form of a relationship (between photographer and photographed, between image and filtering software, between viewer and viewed, between individuals circulating images, between users and social software architectures, etc.).

A selfie is also a practice—a gesture that can send (and is often intended to send) different messages to different individuals, communities, and audiences. This gesture may be dampened, amplified, or modified by social media censorship, social censure, misreading of the sender’s original intent, or adding additional gestures to the mix, such as likes, comments, and remixes.

• The selfie signfiies a sense of human agency…selfies are created, displayed, distributed, tracked, and monetized through an assemblage of nonhuman agents

• The politics of this assemblage renders the selfie—generally considered merely a quotidian gesture of immediacy and co-presence—into a constant reminder that once anything enters digital space, it instantly becomes part of the infrastructure of the digital superpublic, outliving the time and place in which it was original produced, viewed, or circulated.

• Against pathology and moral panics of the selfie

• Selfies as a form of empowerment

• Can speak to social fabric and important cultural moments

• They find that Black and Latino women were relatively comfortable with men of any race taking selfies of any stripe. But “for White women, the ‘man selfie’ must adhere to very strict expectations.” These include the demand that a man’s performance of masculinity seem entirely effortless, natural, and unstaged, lest he come across as less-than-truly-male (or, we might add, less than-truly-White).

• Neither more idealized nor more duplicitous than other sorts of images, selfies invite viewers to think of identity “between the self as an image and as a body, as a constructed effect of representation and as an object and agent of representation.”

(Senft and Byam, 2015) - Authenticity and (Dis) Empowerment

• Authenticitiy:

• For Frosh, a major component of the selfie is the fact that it gestures toward what he calls “corporeal sociability.” As he argues, “the selfie invites viewers, in turn, to make conspicuously communicative, gestural responses” that include taking reaction selfies or responding through gestures such as “like” and “retweet” and “comment.”

• For multiple perspectives to truly circulate material online, issues of authenticity and veracity come into play, as do matters of access and power.

• Senft (2008) uses the term “networked reflective solidarity” to describe the way online networks can link us to others with political sentiments similar to our own and how such solidarity undergirds nearly all viral campaigns on the Internet. Selfies are not necessary for people to engage in networked reflective solidarity. Online political organizing predates photo sharing, social media, and indeed the visual Internet itself. However, when selfies are used, it becomes easier to see how every campaign for solidarity, from the most urgent to the most banal, contains explicit and implicit claims regarding whose suffering and heroism matters, and whose does not.

(Dis) Empowerment

• Yet empowerment also has been narrated through sociological language: To be socially empowered is to have the capacity to take actions in relation to others.

• When speaking of selfies as empowering (or disempowering, for that matter), these registers are worth remembering. The act of creating, uploading, viewing, exchanging, or commenting on a photograph cannot in and of itself lead to any of the states discussed above, though it may facilitate all of them.

• Any time anyone uses a selfie to take a stand against racist, classist, misogynist, homophobic, racist, ageist, or ableist views of what a worthwhile representation is, or should be, issues of political power are clearly at stake.

• One example is the “rubble videos” (Jabari, 2014) in which water-deprived residents of Gaza engage in the ALS “Ice Bucket Challenge” by making videos of themselves dumping sand, rock, and military debris on themselves.

• Such disempowerment might occur online in the form of being disciplined for taking selfies wrong; as racist, misogynist, homophobic, racist, ageist, or ableist attacks; as online bullying; or under the guise of a malicious meme that “borrows” a photo generated for an entirely different audience. It could occur off line, too, as happens when someone finds himself fired at work after being targeted by a revenge porn episode; victimized by doxxing (where personal documentation is hacked and released online); or, worse still, becomes the target of stalking or physical violence

(Senft and byam, 2015) - The Grab

• Social media viewers tend to consume visual material, in a segmented and tactile manner. Grabbing - signifies multiple acts: to touch, to seize for a moment, to capture attention, and to leave ipen to interpretation, rasinging questions of agency, permission and power

• Selfie grabbed as raw data by propriety services and algorithms fuelling features like tickers and newsfeeds. Data is then sold to advertisers through the process known as "mining"

• The fourth level of the selfie involves the grabbing of corporate data by governments and corporations under government contract to aid facial recognition, law enforcement, or profiling efforts.

• Some bodies are grabbed more than others.

• When selfies become politicized, every grab seems to inspire a countergrab. This is especially in the case when photos are seen as overtly manipulative in their aims

• Even when a selfie is remade into a meme for ostensibly good reasons, questions about agency abound. This was the case of the parents of a child with Down syndrome who was alerted that their family photo had been remade into a series of treacly “inspiration” memes (Evelyn, 2014). In these moments, the selfie performs a special of political work, demonstrating three contesting versions of reality at play: the version originally imagined by photographers; the version potentially imagined by anyone using the Internet; and the version where questions of ethics reside.

(Cervi and Divon, 2023) - Overview

• Palestinians use social media as a tool for activism. Each platform provides unique socio-technical affordances that shape users' communicative practices as networked publics

• three memetic templates (#challenge)— (1) lip-syncing, (2) duets, and (3) point-of-view—that unfold the ways TikTok’s design and its play-based affordances ignite affective streams of audiovisual content that render playful activism in times of conflict.

• Driven by TikTok’s culture of imitation and competition, playful activism enables the participation of ordinary users in political emerging events with the help of looping meme videos composed of collaborative, dialogic, and communal socio-technical functions.

•  Playful activism transforms users’ ritualized performances into powerful political instruments on TikTok and makes democratic participation more relatable, tangible, and accessible to various audiences.

• citizen journalism and personal expression allowed evasion of state control over the dominant media narrative, thus democratizing the narra tive by introducing new voices and perspectives and laying the foundations for a new kind of web-based activism

• Online mobility has become part of Palestinian users’ alternatives to a physical lack of mobility, allowing the creation of a new “immaterial territory” (Rousselin, 2016, p. 11) of daily participation that connects geographically remote communities and reconnects Palestinian diasporas.

This cross-territory configuration of medi ated interaction through digital platforms allows Palestinians to reach imagined collectives with whom they engage for civic and political purposes contributing to a counter-visual narration of Palestine

• Testimonial and offer a real-time window into the violence and brutality

•  TikTok’s rich set of vernaculars mobilizes users’ partici pation in many socio-political activities, contextualized in the platform’s playful and humorous cultures

• Playful citizensghip, hashtags as tools for virality

(Cervi and Divon, 2023) - Examples

• Activism through lip-syncing and makeup: makeup tutorials invite users to creatively experiment, mediate, and play with self-representation using audiovisual components. On TikTok, makeup tutorials frequently become activists acts , where "bodies become both the vehicle for action and the action itself"

Face as a canvas for the recontextualisation of the Israeli-Palestinian conflict

• The lip syncing templates on TikTok share qualities of spreadability and legibility and become powerful memetic media (Abidin & Kaye, 2021), allowing users of this challenge to (re)tell the story of Palestinian suffering using a song interpreted by a children’s choir (and thus infusing it with pathos). This song has become a symbol of the melancholic, war-scarred child hood with which most Palestinians identify.

• Participating in trends allows users to act collobaratively

• Musical duets as an Affective activism dialogue: 

• By capital izing on the platform’s tendency to disseminate and expose videos associated with trending challenges (Klug, 2020), users were calling others to Duet with Zein’s version using the caption section on their videos in an attempt to suppress the IDF soldier’s original video. Downplaying the Israeli version and concealing it from the “algorithmic eye” (Abidin, 2020) amplified the virality of Zein’s “hijacked” version of the challenge, making it more visible to random scrollers on the “For You” page. Although users made various memetic performances of Zein’s video, as a public, they communicated in one coherent voice.

Mariam Afifi's Smile of Resistance: POV Challenges as Co-Constructed Empathy

• to Mariam Afifi, a 19-year-old Palestinian activist who became known due to a viral video of her smiling while being beaten and handcuffed by an Israeli soldier in a protest against Israeli authorities (Gill, 2021). Within a matter of hours after her detention by an Israeli soldier, Afifi’s video rapidly achieved viral status, garnering widespread support from hundreds of users who demanded her immediate release from prison. These individuals not only endorsed Afifi’s active participation in the protest, but also helped to amplify a potent and politically charged narrative across a range of social media platforms (see Figure 3a). On TikTok, Palestinians and their supporters transformed her video into a challenge in which they enrolled as re-enac tors of the moment of her arrest, using the POV aesthetic and applying various strategies to convey her presence. In the first variation, users wore dark red hijabs, thereby intertextu ally connecting to Afifi’s hijab in the original footage.

• By playing the role of a victim and integrating photo testimonies that rationalize Afifi’s acts and vehemently disapprove of her arrest “just for defending a girl being beaten by an officer,” users provide global audiences with the chance to transcend the often restrictive framing of the conflict presented by mainstream media outlets

• This playful performance embodies a dialogic infrastructure with the character of the victim because it holds the potential to provide an empathetic view of the suffering of others, even without being the primary source of information about the conflict

• TikTok’s challenges help normalize the idea that political engagement can be an everyday activity or even a play (Zhao & Abidin, 2021), allowing ordinary users novel forms of dialogue around emerging current events by harnessing memetic templates as their political instruments

(Sobande, 2021)

• the anti-Black nature of spectacularized digital (re)presentations of Black people and Blackness, with a particular focus on how racist market logics, racial capitalism, and marketing and branding approaches are implicated in this

• More specifically, due to the pervasiveness of imperialist, white supremacist, capitalist patriarchy (hooks 2000), the ways that Black people’s lives are endangered are not confined to offline settings.

• In turn, I argue for the need for policy-related and media and marketing industry interventions concerning digital racism to account for the wide range of ways that anti-Black racism occurs and is encountered online.

• Anti-Black racism is racism specifically based on the systemic oppression of Black people of African descent and is connected to colourism—the structural favouring and support of light-skinned individuals, including within Black communities, and the contrastingly severe and violent oppression of dark-skinned people

• forms of anti-Black digital racism consistently negatively impact the digital experiences and lives of Black people.

• Black women face a significant amount of online abuse which is often woefully ignored by institution

•  trolls on Twitter have posed as Black women and women of colour to spread disinformation and abuse (see hashtag #YourSlipIsShowing, created by Shafiqah Hudson) (Hampton 2019). Relatedly, research indicates that Black women are 84% more likely than white women to be mentioned in abusive Twitter comments (Amnesty International 2018) and the harm that they face online takes many forms

• Online depictions created by non-Black people, which emulate the physical appearance of Black women in potentially profitable ways, normalisation of the constant sharing of depictions of Black pain and death online , and the anti-Black underpinnings of how Black people’s physical appearance is spectacularized and treated as fodder for the creation of CGI influencers that stem from a white gaze.

 

(Sobande, 2021) - Examples

CGI Black Online Influencers

• CGI racialized online influencers are lifelike digital creations that feature on their own social media profiles which portray them as high-profile, racialized, and marketable micro-celebrities for people and brands to consider engaging with. 

• Overtly racialised due to represening white supremacist representations of blackness

• GI influencer Shudu was created by a white man and was “apparently modelled on the Princess of South Africa Barbie doll”

• “digital canvases that their creators can project messages about so-called racial diversity on, while platforming their own work and without even having to involve or pay real Black people”

• rather endorse people who aren’t even real, rather than paying living black models, this shows that it's strenuous for the representation of black people to be mediated online, even though there have been occurrences of mediation through the black lives matter movement.

• Affects how black identity is digitally depicted and remediated

• Spectacturlaising Black Pain and Lives:

• Media con tent depicting the everyday lives of Black people and their experiences of pain, even, death, is often spectacularized online by individuals and institutions who post, share, remix, reframe and comment on such content in ways embedded in the anti-Black market logics of digital virality, clickbait culture, and the clout that may be afforded to brands that allude to their interest in Black and racial justice work but without substan tially supporting such endeavours

  • constant exposure to extremely disturbing content can have a distinctly detrimental effect on people’s lives and health.

  • the rise of “click-bait culture” which she views as being dependent on distorted, disparaging and discriminatory depictions of Black people, as well as media institutions’ self-serving use of documentation of Black women’s experiences of tragedy, to chase audience numbers and mone

  • Digital Blackface

    • Blackface involves the caricaturing of Black individuals, such as in the form of nine teenth century minstrelsy performances which reduce Black people and their experi ences to racist and stereotypical mannerisms, offensive aesthetic markers and unequivocally anti-Black portrayals. Twenty-first century media discourse on Blackface gathered momentum in 2015, when it was publicly revealed that former NAACP chapter president Rachel Dolezal had been identifying and attempting to present herself as a Black woman for years, despite being a white woman.

    • Certain examples of digital Blackface reflect how some brands’ and non-Black people’s participation in online spaces involves them strategically (re)presenting themselves as being racially ambiguous, and, specifically, falsely alluding to their (non-existent) Blackness—including, at times, as part of attempts to be associated with marketized ideas related to Black social justice activism. Some social media users are manipulating their physical appearance and editing their photographs to appear Black online. In certain cases, this may be due to tacit knowledge of the mar ketability of non-Black individuals appearing to embody an exoticized aesthetic that in some ways may resemble the physical appearance of Black people, especially, Black wome

    • All the same, expressions of anti-Black digital racism are born of deeply entrenched structural oppression that spans centuries and continues to be seeded by imperialist, white supremacist, capitalist patriarchy

 

(D'Ignazio and Klein, 2019)

Bring back the bodies

• Feminism is about power - about who has it, and who doesn't. In a world in which data is power, and that power is wielded unequally, feminism can help us better understand how it operates and how it can be challenged

• Even though we don't see the bodies that data science is reliant upon, it most certainly relies upon them. It relies upon them as sources of data, and it relies upon them to make decisions about data.

• When not all bodies are represented in those decisions - as in the case of the federal and state legislatures which might fund data collection on maternal mortality - well, that's when problems enter in

• Bodies are missing from the data we collect; bodies are extracted into corporate databases; and bodies are absent from the field of data science. Even more, it's the bodies with the most power that are ever present, albeit invisibly, in the products of data science.

• Powerful instiutions…more often than not, control the terms of data collection. Precarious to collect data can put people in harm. Hard to create reliable data e.g. transgender people. It's people and their bodies who can tell us what data will help improve lives, and what data will harm them

• Target Data: worked out when customers were pregannt and sent them coupons… have the resources to collect, store, maintain, and analyse data at the highest levels

• Data capital to consolidate control over customers: values that drive extraction of data represent the interests and priorities of the universities, govts and corporations that are dominanted by elite white men

• Assymetrical relationship between data-collecting instituions and the people about which they collect data

• Bodies are absent from data work: Women are less likely to work in STEM and data collection

• Racial biases in facial analysis technology, biases that already exist in the world are replicated onto these datasets

• Biases will be designed into data systems if the bodies of the system's designers themselves only represent the dominant group

• Bringing back bodies can help avoid data being mined without consent, and bodies that are representative of the population helps to avert the racist, sexist, data products that are inadvertently released into the world

• In the United States, it's white people and their bodies that occupy the default

• To be clear: this does nto mean that there is no value in data or technology. What this means for data science is this: if we truly care about objectviity in our work, we must pay close attention to whose perspective is assumed to be the default.

• feminist projects connect data back to their sources, pointing out the biases and power differentials in their collection environments that may be obscuring their meaning  e.g. Co-liberation instead of accountability, equity instead of fariness, justice instead of ethics

• Just as Benjamin cautions against imagining that data and technology are objective, we must caution ourselves against locating the problems associated with "biased" data and algorithms in technical systems alone

• We should be able to dream of data-driven systems that position co-liberation as their primary design goal

Haraway

• "the god trick." By the "god" part, Haraway refers to how data is often presented as though its inhabits an omniscient godlike perspective. But the "trick" is that the bodies who helped to create the visualisation - whether through providing the underlying data, collecting it, processing it, or designing the image that you see - have themselves been rendered invisible

(Hall, 2013)

• Representation connects meaning and language to culture

• Meaning depends on the relationship between things in the world - and the conceptual system, which can operate as mental representations of them

• we say we 'belong to the same culture' when we interpret the world in roughly similar ways, build up a shared culture of meanings and thus construct a social world which we inhabit together

• Language is a representation involved in the overall process of constructing meaning. Our conceptual map must be constructed into a common language

• Signs: Stand for and represent the concepts and the conceptual relations between them, which we carry around in our heads and together they make up the meaning-systems of our culture

• Can there be effective 'politics of representation'

• 'them' rather than 'us'- are frequently exposed to this binary form of representation. They seem to be represented through sharply opposed, polarized, binary extremes - good/bad, civilised/primitive, ugly/excessively attractive….. They are often required to do both things at the same time

• Every image is also being 'read' in terms of this broader question of cultural belongingness and difference

• Racialised discourse is structured by a set of binary oppositions

• 'civilization' (white) and 'savagery' (black) 'purity' and 'pollution'

• Racial theory applied the Culture/nature distinction differently to the two racilaised groups. Among whites, 'culture' was opposed to 'nature'. Among blakcs, it was assumed, 'culture' conicided with 'nature'. 

• 'Naturalization' is therefore a representational strategy designed to fix 'difference', and thus secure it forever. It is an attempt to halt the inevitable 'slide' of meaning, to secure discursive or ideological 'closure'  

• Stereotyping as a signifying practice

• Stereotyping reduces, essentializes, naturalizes and fixes 'difference'

• Another feature of stereotyping is its practice of closure and exclusion. It symbolically fixes boundaries, and excludes everything which does not belong

• It facilitates the 'binding' or bonding together of all of us who are 'normal' into one 'imagined community'; and it sends into symbolic exile all of them 'the others' who are in some way different

• Hegemony is a form of power based on leadership by a group in many fields of activity at once, so that its ascendancy commands widespread consent and appears natural and inevitable

• Representation, difference and power

• Power… also im broader cultural or symbolic terms, including the power to represent someone or soemthing in a certain way - within a certain 'regime of representation'. It icnludes the exercise of symbolic power through representational practices

• Power not only contrains and prevents: it is also productive. It produces new discourses, new kinds of knowledge

• As Foucualt insists power circulates

• Fetishism, as we have said, involves disavowal. Disavowal is the strategy by means of which a powerful fascination or desire is both indulged and at the same time denied. It is where what has been tabooed nevertheless manages to find a displaced form of representation

• Trans-coding: taking an existing meaning and reappropriating it for new meanings  

 

Case Study AI and Governance

• States are implementing AI Ministers - including France

• Encouragement of AI use and ‘innovation’ for economic growth

• BBC intervew with Liz Kendall - Minister for Science, Innovation and Technology explained how use of AI is encouraged by government officalis and in govt departments

• The New Statesman: ‘the silent coup- How Ai captured Westminster’ AI written legislation

This is a matter of sovereignty. The software products we refer to as “AI” are all built on advanced “foundational models” from the US and China. This is a technology we do not control, but which plays an increasingly active role at every level of the British power structure. It is part of every conversation, drafting emails between officials, summarising ministers’ briefings and composing speeches delivered in the House of Commons.

• Albania: Diella Cabinet AI minister. To curb corruption and betrayal to the govt

• Limited legislation protecting AI and copyright, currently being drafted

• EU AI Act: Banned AI applications in the EU include:

  • Cognitive behavioural manipulation of people or specific vulnerable groups: for example voice-activated toys that encourage dangerous behaviour in children

  • Social scoring AI: classifying people based on behaviour, socio-economic status or personal characteristics

  • Biometric identification and categorisation of people

  • Real-time and remote biometric identification systems, such as facial recognition in public spaces

  • Generative AI, like ChatGPT, will not be classified as high-risk, but will have to comply with transparency requirements and EU copyright law:

    • Disclosing that the content was generated by AI

    • Designing the model to prevent it from generating illegal content

    • Publishing summaries of copyrighted data used for training

(Hall, Heft & Vaughan, 2025)

For social movements essay

Copycats? Do right-wing groups emulate left-wing digital advocacy organisations

• US has has many “copycats”, or cases of direct emulation of a progressive digital advocacy organisation

• Grassfire movement: launched in 2000 with a campaign email petition supporting the Boy Scouts of America’s policy banning homosexual scout leaders. Wanted to reach and mobilisemore than 1million citizens fed up with the “anti-Bush propagana generated by left-wing radical organisations”

• Grassfire has replicated several key elements of MoveOn’s organising model. 1) claims to be grassroots and member driven through digital communications, including social media. 2) claim not to have a massive hierarchy or corporate or organisational structure, goals in to give the tools and impact

• Advance in Australia: Conservative political lobbying group. Relatove success driven by increasing urgency among conservative elites to address the power of GetUP in digital campaigning

• A study of 4 3rd party organsiations in the 2022 Australian election found that advance was the msot active on Facebook and has the most commented-on posts above its left wing inspo GetUp

•  similar tactics to left- wing digital advocacy organizations in particular, through a high number of online petitions but also including media stunts and some offline campaigning. Advance at times also mimics GetUp’s narra tive of member-feedback shaping the organization’s campaigning, such as in a report back to members about advocacy in the 2019 elec tion campaign, where emphasis was placed on the role of members in setting the organization’s agenda

• Alternative explanations: they did not emulate their left-wing counterparts, but adapted to their environment and emulated successful right- wing organizations domestically or internationally.

  • Right messages are local rather than national

(Stier, Posch, Bleier, and Strohmaier, 2017)

When populists become popular: comparing Facebook use by the right-wing movement Pegida and German political parties

• Right-wing populist forces are challenging the established political order across the Western world. Social media has a pivotal role in these processes, enabling populist parties and politicians to bypass media gate keepers and transmit direct mesages to target audiences

• Technological opportunity structures of right-wing social modements liek the Tea Party or the alt-right in the U.S. have also improved significantly since the advent of social media

• the socio technical characteristics of social media make it a unique venue for direct interactions between social movements, their supporters and parties, a core mechanism how emerging societal ideas are established in democracies

Germany

• facebook is well suited for such personalized messages to citizens since it is the social network with the highest societal diffusion (Frees & Koch, 2015). Considering the interactive nature of Facebook, especially politicians who frequently encounter users with populist leanings on their pages might be inclined to adjust their messages

• Pegida: did not create a Twitter account until 2016 . AfD also has a prefernce for Facebook

• higher degree of control that an owner of a Facebook page can exert while debates on Twitter are publicly open and not subject to moderation. In terms of demographics, Facebook is a medium used by a considerable share of the German population, on a daily basis by 22% of Internet users while Twitter use in Germany still remains a ‘special interest.

• This makes Facebook a more attractive medium for populist online com munication. Accordingly, Pegida became increasingly active on Facebook and still attracts significant numbers of supporters online

• There are indications that social media is especially beneficial to movements and parties on the political right, a tendency that has not yet been picked up by the connective action literature. Online social networks allow populists to bypass traditional media gatekeepers and use more radical rhetoric than previously possible in the age of mass media. Right-wing populist communication appeals directly to the people and is par ticularly suited for personalized frames tailored to the interactive user experience: ‘[…] these late modern hybrids invite followers to define “true citizens” as “people like me” (e.g. a white, hard-working native-born citizen) and not those immigrants who come to live off my hard-earned tax money’

• The narrow focus on social media use by movements from the political left needs to be reconsidered and synchronized with established research on populism

• Pegida openly advertised AfD contents on its Facebook page. We thus (1) expect that the AfD attracts the highest share of Pegida supporters and (2) that the AfD emphasizes populist issues in order to attract the voters with preferences similar to Pegida activists.

• In a counter factual situation without the emergence of Pegida, which revealed the substantial reson ance of right-wing positions in German society, the more nationalist forces in the AfD might not have prevailed in the internal party struggle during 2015