Algorithmic vulnerabilities and the datalogical: Early motherhood and tracking-as-care regimes

Context

The contemporary baby tracking market is relatively buoyant. Not only has tracking culture and the datafied self become increasingly normative in the United Kingdom (see Lupton, 2016), but there is also a longer term and more historical tracking-as-care discourse of early motherhood (see Wajcman, 1991) which has shifted to the digital where it is also part of the mhealth (mobile health) movement in the United Kingdom (Speciale and Freytsis, 2013). Tracking and biosensor apps occur within the intimate and neoliberal frameworks of ‘knowing yourself’ better (Lupton, 2016; Sanders, 2017). They are highly personal, often uneven and malleable data sets that rely on automatic, biosensor and inputted information to varying degrees, and as such they are negotiated differently, played with and appropriated. During the course of this project, the women used a range of apps – starting with BabyFeed and Baby Buddy where data are inputted manually, and moving (or returning) at around months 4 and 5 to self-tracking apps – health monitors and pedometers such as Jawbone and Fitbit, Garmin & Apple watches. Some of the mothers used baby-weaning apps like Made for Mums for a few weeks between 4 months and 6 months, although there was much less take up of weaning apps within the project. Tracking and biosensor apps offer users and interested parties ‘granulated’ data of ‘pixelated’ selves (Topol, 2012: 231). As Kitchin et al. (2016) and others (Day and Lury, 2016; Neff and Nafus, 2016) have noted, such apps atomize us, and in so doing are part of a long history of the atomization of – particularly female – bodies which has occurred mostly through health and medical discourses as well as within wider popular culture (Thornham, 2018). For those of us that use them, they enable us to find, as Greenfield argues, ‘personal answers to personal questions’, seeking a ‘different kind of knowledge’ by comparison with traditional expertise (Greenfield, 2016: 131–132). They are intimate, individual and are built on and through a neoliberal discourse of individualism, choice, meritocracy and mobility. Tracking and biosensor apps are part of a wider discourse of biopower, disciplining bodies over time while also generating data to be used by others to discipline us, thus reconfiguring traditional notions of structure and agency (O’Riordan, 2017: 49).

Tracking apps, wearables and biosensors are interested in very particular data, producing dashboards that are necessarily meaningful on a range of levels, for a variety of stakeholders. And although they may be highly personal, they are also, as Deborah Lupton reminds us, inherently tied to, and the product of, ‘broader social, cultural and political processes’ (2016: 1). It is worth noting, for example, that Baby Connect is developed by Seacloud Software, a mobile app developer company in California, USA. Any data generated outside the United States are transferred and stored in the United States in keeping with their data protection and privacy laws. Seacloud Software collects usage data (click data, duration, time and date), technical data (IP addresses, devices operating system, geolocation, browser language) as well as the inputted data ‘itself’ regarding length of sleep, feeding patterns, growth, gender and so on, of the infant. Usage and technical data are shared primarily with subprocessors for the app: Amazon Web Services, Google, Groove, Twilio, Recurly; but Seacloud reserves the right to share anonymized data inputted by individuals for business or research purposes. ¹ It operates along the lines of what José van Dijck has called a ‘cultivated garden politics’ (2013: 164) whereby the data provided by the app may be sold in carefully packaged ways to other – usually health – organizations. Baby Feed (the app most of us used) is produced by a software development company Fehners Software LLP (Haywards Heath, UK) and has a sliding scale of data privacy insofar as the paid version only enables subprocessors to collect usage and technical data from users and only via synching processes with other devices (apple watches, garmins, other iPhones): The company itself claims to collect no data whatsoever. The free version, however, includes the Google AdMob SDK plugin, which means that data are governed by Google data policies which are much more open (collecting GPS, sensor data, other services in the proximity of the device such as Wi-Fi access points and other Bluetooth-enabled devices). ² Baby Buddy, by comparison, is owned by the charity Best Beginnings and was developed as part of the shift towards mhealth in the United Kingdom within the third and public sectors (see e.g. Becker et al., 2014; Sama et al., 2014) and was funded in the final stages of development by the Big Lottery. It was designed in conjunction with Guys and St Thomas’ and Blackpool’s NHS Foundation Trusts where it was embedded into their service delivery as part of the pilot testing period. ³ The rationale for the app was to reach parents who are ‘least likely to engage with traditional forms of health information’ (Cooper, 2015: 7) but was prompted in its design by the ‘advent’ of mhealth interventions (ibid.: 8) and by comparison with Baby Connect and Baby Feed incorporates an appointment diary and ‘what to ask’ functions for interactions with health professionals. Although the three apps have very different data policies and origins then, they have comparable functionality, dashboards and data interests as I will go on to explore and indicate quite clearly (I think) the range of data cross over not only between the private and public sectors but also between the consumer and healthcare interests.

Datalogical perinatal care

Baby Connect, Baby Feed and Baby Buddy are part of a number of health apps endorsed – in their advertising first of all, but also discursively, and as part of perinatal professional practice (as with Baby Buddy) – by health professionals. As tracking apps, they following in the footsteps not only of the health apps and biosensors discussed most prominently in the literature on datafied and quantified selves (see e.g. Lupton, 2016; Nafus, 2016; Neff and Nafus, 2016; O’Riordan, 2017) but they also more overtly follow the ovulation and period tracking apps used for conception and widely discussed on forums like netmums and BabyCentre. Increasingly (as with Baby Buddy), such apps are not just pre- or postnatal but perinatal, incorporating conception, pregnancy and postnatal experiences in a single app thus prolonging their use and collating much more nuanced data. For our project, a discussion of postnatal and perinatal apps came up in the very first conversations we had and when I later asked in the first month of our project how the women came to know about apps like Baby Connect, Baby Feed and Baby Buddy they detailed a range of discussions with midwifes friends and family as well as online forums and platforms:

my midwife suggested some names [of apps] I think coz I couldn’t remember how long she’d fed for one time, but I knew about like Made for Mums coz my sister uses it and I used Glow ⁴ actually when we were trying to conceive. (Sanja, month 1)

it was actually my sister what suggested Baby Buddy. She said it were really good for like, keeping records, seeing what to do and ask. (Naveena, month 1)

They [midwives] gave me this pack with all this info in – this list of apps and resources and that. But I’ve got a period tracker app and a FitBit, so its second nature really. (Aimee, month 1)

I was my midwife. I think she just said it was a useful thing. Like, as in, you know, other mums are using this. I think she told me, is it Baby Buddy? That one. (Lee-Ann, month 1)

Baby Connect, Baby Feed and Baby Buddy also monitor these data – the length of feed, the weight of the baby, the side of the feed, time between feeds: Baby Buddy monitors the latch and position of the baby during feed as well. Unlike the apps though, such data within the postnatal healthcare system that is gathered at every appointment between mother, infant and health professional are constitutive, algorithmic and based on a series of gatekeeping (‘yes/no’) decisions through questions such as: is the baby ‘gaining weight?’ does the baby feed ‘for 5-30 minutes at each feed?’ ⁷ Such questions form part of a decision-making process around which medical and health intervention is based. Women are routinely asked questions pertaining to such data at each appointment, which is of course one reason why such apps are meaningful, but not knowing the answer, resisting the question or offering the ‘wrong’ answer, generates a whole range of subjective anxieties and annoyances as well as interventions as the quotes below detail. The answers mothers give to health visitors and midwives generate processes and procedures within the very regulated and metrified system of care, a system that itself feeds into wider health and social care systems. The women know this, not least because as I said at the start of the article, we met at a breastfeeding clinic to which we were all referred precisely because of our breastfeeding metrics:

[My midwife] suggested I came – I’m really struggling. She’s lost 9% of her body weight. I’m trying to feed her every 2 and half hours, but it’s really difficult. (Karen, month 1)

She’s underweight so the midwife referred me. I’m not feeding her right. (Naveena, month 1)

I needed to get out of the house. Simple as. I’ve not been out in 2 days. I haven’t slept so they sent me here. (Shelly, month 1)

Cracked nipples, mastitis, cysts. Engorged boobs, leaking nipples, bleeding nipples. You name it, I’ve got it. (Akira, month 1)

Indeed, as Lee-Ann says in the excerpt above, it is the app that tells her what she is doing ‘okay’ and what she needs to change as a mothering practice. Naveena tells us that her app underpins her meetings with her midwife, and she uses it to demonstrate her mothering practices. More than this though and as suggested above, it is the data which is trusted as opposed to her own experience: ‘she doesn’t just have to take my word for it’, conceiving of data as offering ‘certainty’ which are set against her own ‘untrustworthy, inexact’ expertise (Lupton, 2016: 94). Lupton argues that we increasingly trust ‘data over embodied knowledge’ (2016: 94) in relation to apps and wearables, while Kate O’Riordan has talked recently about how data are becoming material through such apps because of what is valued, researched and measured (2017). The fact that mothers also, in their conversations in the breastfeeding cafe, echoed these metrics in their own discussions about their baby’s ‘progress’ is indicative of the power of these metrics. There are a number of issues to note here then, firstly in terms of the way data are valued and conceived as more accurate than the lived experiences of maternal subjectivity – and issue I will return to below and which has resonances with a long history of medicalizing childbirth, pregnancy and early motherhood. This suggests at the very least a continuity in terms of the ideological underpinnings of both maternal subjectivity and indeed data rather than a shift prompted by wearables and biosensors. The second issue to note relates to the way the metrics discussed during medical appointments with midwives and health professionals bleed across into other spaces to become markers of successful mothering per se. The fact that the women constantly talked about how many feeds their infant had in every 24-h period; the length and quality of feed; and whether the baby fed from both breasts or one, doubly emphasizes not only the power of these metrics, but also the way they become enmeshed in subjective and normative markers of successful mothering. What is valued about motherhood, indeed what motherhood is – how it is experienced, discussed, measured, is ‘leaking’ across (to borrow Cheney-Lippold’s term 2017: 143) from the data and algorithmic to the subjective and vice versa. This is precisely what Fotopoulou (2016: 96–99) argues when she suggests, drawing on Foucault’s concept of biopower, that wearables and biosensors ‘discipline’ the body through a lived negotiation with regulatory power systems and processes (see also Lupton, 2016: 1; Sanders, 2017: 40). This means that, at the very least, the data recorded by the apps, along with the processes of inputting data by the women, are far from benign and should be regarded as part of a process of soft biopolitics – of conditioning into particular structures of surveillance and care as well as wider ideologies of maternal subjectivity.

Motherhood, metrics, agency

The apps work by atomizing and disaggregating lived experiences of motherhood into metrics that can then be re-aggregated divest from the messiness of lived experiences and in ways that can be plotted against standardized data visualizations. This atomization and re-aggregation is a familiar process of medical and healthcare well-documented by scholars (such as Latimer, 2013; Mol, 2002; Mol and Law, 2004) in terms of the simultaneous objectification and pixilation of bodies – a process according to these scholars that ultimately renders subjects powerless (and vulnerable) on many levels. In drawing these correlations, there is also a familiar feeling of déjà vu here and I am thinking of long-running feminist concerns with the medicalization of childbirth and early motherhood as well as the atomization not only of the female body per se in wider cultures (see e.g. Gill, 2007; Pollock, 1987) but also within wider medical discourse. Lucy Wajcman tells us that technologies ‘coexist with a powerful ideology of motherhood’ (1991: 57) and have long sought to atomize complex lived relations and ontologies into discrete and separate data and technological processes. Her historical work on the relationship between technology and motherhood highlights not only the longevity of this process of the atomization of the female body through technological processes (1991: 70 see also Michael and Rosengarten, 2012; Shaw, 2012) but also demonstrates a long-term failure to conceive of motherhood in and of itself as a constitutive category within systems. Instead, motherhood is defined as the outcome of aggregated data, like in these apps: How motherhood is subjectively understood and valued by these women is in part constituted through these metrics and legitimated by them. As many feminist scholars have noted (see e.g. Crowe, 1987; Franklin, 2010; Lawler, 2000; Minden, 1987; Shaw, 2012), this is long running and among other things, works to cement the notion of technology as validator and legitimator of motherhood and maintains a conceptual separation between technology and maternal subjectivity. As Wajcman reminds us, ‘machines inexorably direct the attention of both the doctor and the patient…towards the measurable aspects of illness’ (1991: 70). This means that technologies (e.g. scales, syringes, blood pressure monitors, apps) and metrics have long been conceived of as more ‘honest’ as N Katherine Hayles discusses (2017: 126–127), truthful and accurate (Gitelman and Jackson, 2013: 2) than lived embodied experience and that this is part of a wider datalogical environment that constitutes maternal subjectivity. What I am arguing then, is that valuing data over lived experience is not a new phenomenon, nor is the atomization of female and maternal bodies and so on many levels the wider discursive and every day normativity of these apps should be no surprise. But this also means as Fotopoulou and O’Riordan have argued in relation to biosensor apps, there is much more ‘at stake’ here, (2016: 2) than we might realize.

Algorithmic vulnerabilities

In the second half of this article, I want to consider what is absent in the apps and wider health metrics in terms of data silences. Most notably, I will discuss the lived embodied experiences of early motherhood which are entirely absent and which generates a range of subjective anxieties and vulnerabilities for the women of this project as they navigate their own experiences within the frameworks of the wider datalogical healthcare provision in which their apps are complicit. More than this though, I argue that the fundamental premise of tracking and biosensor apps is at odds with maternal subjectivity as discussed by feminist scholars. What this means in thinking about the datalogical and indeed digital culture more generally is that there is huge silence if not negation of lived experience and maternal subjectivity that is increasingly infrastructural, cultural and problematically normative, partly because they feed and perpetuate a wider ‘taken-for granted’ sensibility of gendered neoliberalism (Gill, 2017: 609) which I will go on to explore.

As Frankie tells us, ‘what all these nice graphs and charts don’t tell you, is how much it fucking hurts!’ (Frankie, month 1). Indeed, the apps do not count or measure pain, frustration or anxiety. There is no way to chart unsuccessful or aborted attempts at bottle or breastfeeding, for example. Similarly, the apps count intentional breast or bottle-feeding only – not unintentional leakages; they count the duration and frequency of sleep, not the quality. Although nuances and complexities can be added as a note to inputted data, this functions in a non-constitutive way – as bounded or ‘stranded’ data (Singh, 2012) – information that is also devalued here because such notes are a complication and messiness to a clean and simple, ‘scientific’ and atomized metric. Such silences are not just evident in the data (or lack of it); they are in the design and aesthetics of the apps too. The ‘clean and proper’ (to borrow Kristeva’s term 1980: 102) design of the apps, their pastel colours, simple 2-D geometric shapes and overall uncluttered aesthetic renders invisible – indeed, erases – the ‘fleshy’ maternal body discussed in much feminist scholarship (Battersby, 1998: 11, Shaw, 2012: 121) and related subjectivities. These are dashboards that explicitly make breastfeeding and the lived experiences of it entirely invisible. Maternal subjectivity is not a concern for the apps: data are. In recognizing this, we also need to note that this is entirely in keeping with wider discourses around motherhood and maternity, as well as the discussions above in terms of the various interactions with healthcare professionals and their interest in certain data. For scholars such as Sarah Franklin (2010) what is important about such long-running discourses is the generation of an ultimately patriarchal epistemology – which arguably is exactly what is represented in these apps. For other scholars such as Bassett (2010) and Firestone (1979), what is important is the obscuring, if not ‘obliteration’, of the mother within medical discourses (Bassett, 2010: 97). My suggestion in thinking about these apps is that we can find resonances of these concerns here too – in the silences of the data and in the design of the apps. Maternal identity is firmly located at the site of embodied experience: bound up in the lived practices that may be partly constituted by the apps, but are not located or visible within them. Thus, the silences and politics of the app are entirely normative in terms of wider discourses of maternal subjectivity in which we can trace a strong and consistent thread of the negation or eradication of the mother. As Akira tells us:

What I really want is something on here that also counts how many times I have been vomited on, shat on, peed on. I swear, at least 3 times a day. Then I could be like, ‘well he hasn’t fed 8 times in the last 24 hours, but he HAS puked all over me 10 times – is that ok?’. (Akira, month 1, 2016)

One contributing factor to the disappearance or silencing of shared and common lived experiences might be found in the way the data for wearables and biosensors work. Indeed and by comparison with traditional medical data, which works by statistically verifying an individuals data by correlating them with bigger (big data) real-time data that are in turn correlated at a range of points (for example age, gender, ethnicity, medical history as well as similar stage of illness); wearables and biosensors work with what Greenfield has called the ‘n of 1’ principle (2016: 125):

The n of 1 rejects the requirements of large numbers of subjects for statistical validity and expert credentials, forging a new epistemology of health and being where the single case or person collecting the data over a lifetime displaces the population as locus for knowledge and intervention. (2016: 125)

Perhaps, more importantly, the n of 1 principle only works if we believe in a whole, unified and consistent self over time and space. The data of wearables and biosensors are only ever meaningful if we assume that the constant factor that can be metrically assumed is a whole, unified and complete subject. But maternal subjectivity, as many theorists have noted (Battersby, 1998; Baraitser, 2009; Tyler, 2009), encompasses a range of embodied subjectivities and relations (pregnant body, birthing body, mothering body), each also bound up in wider ideologies of embodied female form with its inescapable and inevitable promise of ambiguity, contradiction or mutation through its reproductive capabilities (see also Thornham, 2015). It is in fact at the opposite end of the spectrum to a whole, unified consistent self over time. The apps emphasis on certain metrics – such as latch of baby, position of baby while feeding, whether both breasts are offered – is logical not least because validity through consistency over time is problematic. But the inflexibility this produces in terms of logging the experiences of maternal subjectivity creates real data silences which contribute to an overall feeling among the women that such experiences are devalued, unimportant and undermined. We can see this sentiment in the comments from Frankie and Akira quoted above but also in the following discussions which detail conversations about cracked or bleeding nipples and mastitis with medical professionals:

she asked me how often the baby was feeding and when I said how sore it was, she told me to just ‘feed through the pain’ and my nipples needed to ‘toughen up’. I cried when she left. Honestly. I just cried and cried. (Aimee, month 1)

I went to the out of hours surgery coz of this lump which was so bloody sore. Couldn’t touch it, I like thought it were mastitis, but weren’t sure – like you don’t know do you? So the doctor asked me how old [baby] was and then just wrote me out three-week prescription for antibiotics. Never really looked at me boob, never really asked me much, just said the pills were okay if I were breast-feeding. I wanted to like talk about it, and I was in and out before I could take a breath! (Mae, month 3)

Conclusions

Wilson and Yochin in their book Mothering Through Precarity (2017) talk about how the practices of motherhood are increasingly becoming understood in atomatized ways and also becoming overly valued as a marker of successful or competent mothering. They don’t talk about data and apps specifically in their book but are instead concerned with what they call the ‘digital mundane’: the lived and everyday practices of motherhood and how these mundane practices give sense and meaning to peoples lives (2017: 3). What they do discuss, however, is how everyday discrete decisions, for example, about length of sleep for the baby, amount and type of food, breastfeeding, clothes, playtime activities are all being atomized and simultaneously elevated to the consequential. They argue that motherhood is increasingly being conceptually formed through the aggregation of discrete and atomized measurable elements, and the critical point they make is that maternal subjectivity is being ‘liberalised’ (Wilson and Yochin, 2016: 50) in a self-perpetuating cycle that is affectively powerful. Elevating the atomized, pixilated, discrete decisions of motherhood to the consequential – allowing them to become constitutive data in the wider discursive meaning of motherhood – constructs mothers as always responsible, accountable and productive. Mothers are ‘important gender citizens’ with ‘weighty social responsibilities’ (2016: 50), but this means the happiness of the family rests on not just the ‘minutiae of mothers’ actions but also and importantly ‘their broader lifestyle choices’ (ibid.) And as Oulette and Wilson have argued, these choices are part of a refashioning of motherhood to make ‘the feminised labor of caring for others more compatible with the self-enterprising ethos demanded by today’s neoliberal policies and reforms’ (2011: 55). The combined argument of these scholars is that that the digital feeds and perpetuates what we might call the gendered neoliberalism of contemporary motherhood (to borrow Gill’s phrase 2017: 609). For these scholars, the digital does this through an endorsement of resonant politics of neoliberalism, choice, meritocracy and enterprise, which all contribute to emotional anxieties and feelings of inadequacy in a perpetual cycle in which the women are emotionally vulnerable to and affected by gendered neoliberalism that is always-already algorithmic and digital. There are no cracked and bleeding nipples, no shit or vomit-covered mothers within ideologies of self-enterprising individualism or mother-as-citizen. To reiterate Day and Lury’s phrase, this is a recursive fractal, not (only) of data, but of gendered neoliberalism framed within the resonant discourses of motherhood.

Bassett has argued that what has changed in the era of the quantified self is that bodies and computational devices ‘intertwine to measure the human day and co-constitute the world in which we live’ (2015: 136) and if we think about maternal subjectivity in relation to how it is being co-constituted, we find an overriding gendered neoliberal discourse which is atomizing not only maternal bodies, but practices, subjectivities and relations. This atomization may well be long-standing but – as Bassett (2015) and Cheney-Lippold (2017) also remind us, it is the technological and the datalogical that is gaining traction not only in terms of decision making power, policy, normative practices and discourses – as you can see from the quotes here – but also in terms of our own conceptions of our own identities. Quite often, as Grosz has argued, this is happing at the expense of lived experience (2001), and for the women of this project, the consequence of an increasing irreconcilability between lived experience and the datalogical construction of it is manifest in real anxiety about doing motherhood ‘wrong’. It is manifest the double silencing across the datalogical and the everyday of certain experiences and data – like cracked and bleeding nipples, like oscillating emotional states, like changing bodies, opinions, preferences and priorities.

It is important to say that I am not suggesting the women quoted in this article do not resist or question such positionings. Their perceptions and experiences of the apps changed over the 6 months of the project, as did their willingness to use or engage with them:

I don’t mind pacing in the night – I think ‘that’s another 500 steps’ [laughs]. And in the middle of the night, you can like look and go, ‘over 10,000 steps at least. Good.’ I mean I know this is the point of it, right? But I still take it like a little reward. (Naveena, month 4)

I realised I was reaching to turn on the app before I started feeding. There I was, she was screaming, and I was scrabbling for my phone! I thought ‘what am I doing here?! This is ridiculous!’ (Lee-Ann, month 5)

It showed me that I was feeding longer on the left than the right, like 3 minutes more every feed, so then I changed how I fed and then it was the same. And then, looking back I think, ‘why did I do that? Did it really matter?’ I dunno (Karen, month 6)

What I am suggesting here is two things. The first is that the disciplining process of the apps occurs through mundane and routine behaviour – which in itself assumes some sort of reciprocity between the app and the wider maternal experiences. Tracking-as-care, food diaries, calorie intake, pedometers, for example, are long-term and familiar beyond the specificities of the quantified self or mhealth ‘movements’: Being data is a familiar lived experience. For the mothers of this project, for example, there was a deep correlation between ideologies of motherhood that were interwoven with the ideas of tracking-as-care, the intensified surveillance of women (Winch, 2015), the long-standing (and exacerbated through discourses postfeminism as Gill argues 2017: 607) ‘need to monitor and discipline one’s self’ (Ouellette, 2016), the metrics demanded by the wider datalogical health care system as markers of health and their own uptake of baby monitoring and fitness apps.

My second suggestion is that these practices do not exclude a conscious consideration and even critique of what is occurring: These are not, then, the ‘unreflected, taken-for-granted’ actions that Shaun Moores talks about in relation to embodied and tactile digital relations (2014: 202). We can read this in two ways. We can see it as opening up a potential space of resistance – in keeping with long arguments within the quantified self community (see Lupton, 2016, Neff and Nafus, 2016) – whereby possibilities of resistance and the hacking of these processes are enabled precisely because the politics of the apps become/already are visible. The visibility of the politics combined with a critique of them is what potentially generates a space for resistance, reappropriation or disruption – enabling such apps to be ‘turned inside out’ like Greenfield’s medical data (2016: 126) and perhaps go ‘well beyond the individual’s quest for self-knowledge and self-improvement’ (Lupton, 2016: 143). This perhaps would prompt the suggestion that the women are less algorithmically ‘vulnerable’ than I am proposing, not least because, seen here, they would have more negotiating agency.

The second way of reading these actions is less straightforward. Nafus has talked about how tracking apps encourage a process of reflection, making us ‘think twice about the social relations [we] believe [our] bodies to be in’ as well as the ‘materials that constitute both those bodies and their ecosystems’ (2016: 228). And the comments above – and indeed throughout this chapter – can be read as part and parcel of such reflections. Many of the women during the project articulated concerns about self-tracking and baby monitoring, but everyone continued to self-track on some level, and this means we have to question the power of reflection within a wider disciplinary process. In thinking about processes of disciplining (Foucault, 1977: 221) or ‘automata’ (ibid. 136), there is something really interesting here not only about the incremental reach of productive labour or work but also about the conscious awareness that sits alongside – sometimes unproblematically – the politics of the apps that makes them demanding. Indeed, Cheney-Lippold argues that even though algorithmic identities can never ‘truly square’ with our lived experiences (2017: 145), they are nevertheless gaining traction in terms of decision-making power, policy, normative practices and discourses. Agency or autonomy is not with the body in this scenario, but with the technological economy that reconfigures the body in its own terms.