Intersectional observations of the Human Brain Project’s approach to sex and gender

Introduction

The Human Brain Project (HBP) is an enormous research endeavour with immense transformative potential within information and communication technology (ICT), neuroscience and related fields. It was funded in 2013 as a Future and Emerging Technologies Flagship to increase understanding of the brain from diverse intellectual perspectives and develop technical and ethical governance infrastructure for the advancement of neuroscience.

Like many ICT projects, women are underrepresented in the HBP. The proportionate lack of women in the HBP and similar initiatives is an ethical issue, not only because consideration for sex and gender throughout the process of research is vital to producing high-quality science (Nielsen et al., 2017 for a recent review), and because multiple perspectives are important in producing socially relevant research, but also for reasons of fairness and social justice (Monroe et al., 2014; Salinas and Bagni, 2017). The need to create truly diverse teams in ICT is immediate. Some emergent technologies are already ethically problematic, as demonstrated by the misclassification of Black women, transgender people and non-binary people by automatic gender recognition (Buolamwini and Gebru, 2018; Keyes, 2018). Engineering novel techniques and technologies for the betterment of society is central to the HBP mission, which makes representation at every level a particularly important consideration; thus, actions on these issues should not be taken for purely instrumental reasons.

The underrepresentation of women in the HBP has been acknowledged as a serious challenge within the leadership of the Project, who have progressively implemented an approach to the issue. However, the methods by which representation of women is achieved are worth reflecting upon, because some tactics may backfire and reinforce negative trends (Roberts, 2015). This could mean a significant loss of investment and a failure to increase diverse representation, thus robbing the project not only of prospective talent, but of public trust and potential social relevance. In this paper, we critically assess how approaches to sex and gender in large technical ICT projects are realised by using the HBP as a case study. We will contextualise the strategy of the HBP within the wider context of the representation of women in ICT, and reflect upon this approach from an intersectional standpoint.

Our intention with this work is to offer insights on the HBP strategy which may be helpful for both the HBP and future ICT projects; contribute to wider discussions on multiple inequalities and intersectionality within an EU policy framework; and raise awareness of the risks of focussing on only one axis of inequality in attempting to increase “diversity” in technical ICT projects.

Definitions and Sources

We follow Johnson et al. (2009) in our definitions of sex and gender because these were designed for biomedical research and incorporate theoretical aspects derived from the humanities and social sciences:

Sex is a multidimensional biological construct that encompasses anatomy, physiology, genes, and hormones, which together affect how we are labelled and treated in the world. Although conceptualizing sex usually relies on the female/male binary, in reality, individuals’ sex characteristics exist on a fluid and medically or socially constructed continuum. (Johnson et al., 2009, p. 3) [emphasis ours]

Although sex is primarily biologically constituted, it is not independent of social construction nor is it binary. When explicit, we use the term “female” to mean a person identifying as a member of the female biological sex, and the term “male” to mean a person identifying as a member of the male biological sex. In the observations described here, we have mainly perceived gender or descriptions thereof

Gender is a multidimensional social construct that is culturally based and historically specific, and thus constantly changing. Gender refers to the socially prescribed and experienced dimensions of “femaleness” or “maleness” in a society, and is manifested at many levels. The experience of gender is always linked to the social and political context. (Johnson et al., 2009, p. 4) [emphasis ours]

Gender is not universally understood in the same way and is not static; for example, consider the complexity inherent in geographical regions such as Europe (Verloo, 2007). Like sex, gender is also not binary and is deeply performative (Butler, 2006). We use the terms “she/woman” and “he/man” to describe these genders only when they are explicit. If not, we use “one” or the singular “they”. Sex and gender are not interchangeable, and the intrinsic linkage of both to wider cultural contexts underlines the importance and potential value of an intersectional approach.

Intersectionality was defined by Shields (2008) as “the mutually constitutive relations among social identities” (p. 301), and Walby et al. (2012) state that “inequalities mutually shape each other […] […] at their point of intersection” (p. 237). Both definitions serve our purpose here, and we agree that every person concomitantly possesses multiple identities and aspects which shape their experiences (particularly their experiences of oppression), and that these cannot simply be individually disentangled from the others. Identities, categories of social membership such as ability, class, gender reassignment, age, religion, race or ethnicity, sexuality, etc. are not easily separable, and to treat any single aspect in isolation is to contribute to inequality and oppression. The example par excellence is provided by Kimberlé Crenshaw (1991): “the intersection of racism and sexism factors into Black women’s lives in ways that cannot be captured wholly by looking at the race or gender dimensions” (p. 1244). In that context, both discourses of anti-racism or feminism marginalise Black women because they privilege the views of Black men or White women respectively (see also Collins, 2008).

These definitions inform our discussion of representation in ICT, the HBP’s approach to “Gender and Diversity” through to March, 2018, and our reflections on how this strategy relates to other policies and developments. We will contextualise the HBP’s situation within ICT more broadly, and discuss observations of the available materials and events relating to the HBP’s approach.

Methodology

We have written this paper from the perspective of researchers within the HBP. There are benefits and challenges of writing a pseudo-auto-ethnographical account of a project within which one is embedded. Usefully, we have been able to access the Project Blue Book (high-level personnel directory), an HBP conference on gender, a related workshop, internal circulars and advertisements and PowerPoint slides as members of the HBP. Phenomenological aspects of these events and items have heavily influenced this paper, and we have personal knowledge of circumstances (e.g. knowing approximately how many people work in the HBP; advice that the Gender partner was changing). However, our critique is tempered by an awareness of our positions within the project, and made in full knowledge of the importance of avoiding alienation of our colleagues and collaborators.

We intended to examine the HBP Gender Survey, sent to the entirety of the Project in Spring of 2017 and the HBP Gender Action Plan (GAP), to conduct a comparative analysis of baseline data with the data on underrepresented groups reported elsewhere (Napier Diversity Report, https://epsrc.ukri.org/newsevents/pubs/napierdiversityreport/). At the time of writing, that version of the GAP was not available within the HBP, and we were not provided with the survey or its methodology. Thus, aspects of the HBP strategy discussed here are based upon first-hand experiences and observations of the presentations used at the Gender and Diversity Conference (9 March, 2018) as well as the Career-Building Workshop (8 March, 2018) and related discussions with speakers and participants at both events. We incorporate observations of textual materials relating to these events and “gender”-related HBP Open Calls. Additionally, we consulted policy documents relating to the Horizon 2020 research framework.

Representation in ICT

At this time, women are largely underrepresented within ICT education and practice in North America and Europe (Nedomova and Doucek, 2015; Pechtelidis et al., 2015; Sax et al., 2017; though see Varma and Kapur (2015) for India as a contrasting example and Wakunuma (2007) for the case of Zambia). A litany of books and articles from the previous decade traces the problematic experiences of women in computing education and related disciplines (Fisher and Margolis, 2002; Henwood, 2000; Papastergiou, 2008; Cheryan et al., 2009; Misa, 2010). This mirrors issues of representation in academic leadership (Monroe et al., 2014), particularly in Science, Technology, Engineering and Mathematics (STEM) disciplines, and supports the case for considering representation in computing independently (Sax et al., 2017).

Initiatives intended to increase the proportion of “women and underrepresented minorities” in STEM and ICT are viewed as a multi-purpose solution to issues of specialist labour shortage, a way of fuelling innovation or as a method of shaping a more diverse, representative future (Roberts et al., 2002; Lagesen, 2007; Henwood, 2000; Bosch, 2015; Rodriguez and Lehman, 2017). There are many complex cultural, systemic and infrastructural factors contributing to the underrepresentation of women in these areas, including the early age at which activities may be gendered and the pervasiveness of negative attitudes toward women in certain professions (Pearce, 2017). This has resulted in many interpretations of the core nature of the problem and multiple framings of women. In many of these cases, women are presented as a homogenous group posing a problem to solve (Henwood, 2000), the answer to problems of “equality” (Monroe et al., 2014; Salinas and Bagni, 2017) or as a way of improving research and innovation (Nielsen et al., 2017).

Published articles suggest strategies to improve the inclusion of women, ranging from ways to achieve “gender equity/equality” at scientific events and conferences (Débarre et al., 2018; Moghaddam and Gur, 2016), to lists of policies or actions to implement (Monroe et al., 2014) to picking apart the many contributing factors when women select (or exclude) ICT degrees or professions (Sax et al., 2017), to simply concluding that since the numbers of women in ICT roles are rising overall, that the problem with fix itself (Nedomova and Doucek, 2015).

However, a diverse, representative workforce with the capacity to produce the desired trends in innovation cannot be achieved by simply “hiring women”, implementing “family-friendly” policies (Monroe et al., 2014) or even addressing issues of stereotyping, identity dissonance and individual belonging (Henwood, 2000; Bosch, 2015; Pechtelidis et al., 2015; Rodriguez and Lehman, 2017). People hold multiple categories of social membership (identities) concomitantly (Museus and Griffin, 2011), and these mutually shape each other and contingent social relations (Walby et al., 2012). Therefore, attempts to achieve “diversity” solely through “gender” are problematic because there is no such thing as “a woman”: one’s identity is multivariate and fluctuates. To focus questions about inclusion on a single variable (in this case, sex or gender, though these are frequently conflated) can exclude groups of people, especially when other aspects such as class or “race/ethnicity” are taken to be neutral or default categories (e.g. “whiteness” following Carbado, 2013). Efforts to increase the number of women in academia, STEM or ICT tend to focus on “women”, are often not intersectional and can thus serve to further marginalise those who are not in positions of privilege to begin with (e.g. women and non-binary people who are not White, able, middle class, cis-gendered, etc.).

Intersectional methodologies are incorporated into research design and analysis (Windsong, 2018), and Rodriguez and Lehman (2017) advocate for an intersectional agenda in ICT, drawing on decades of feminist and critical race theory research. Kimberlé Crenshaw’s seminal text detailed how the experiences of being a Black woman are not simply a combination of experiencing being Black (with the idea of “man” as default) and experiencing being a woman (with the idea of “White” as default; Crenshaw, 1991). Black women and Black LGBTQ academics in computing experience an intensely inhospitable environment (Payton et al., 2018). Harris and Daniels (2017) note the hostility experienced by Black lesbians in the tech industry, and Gray (2012) describes the oppression of Black and Latinx sexual minorities in digitally mediated spaces. Religion also affects whether women consider a career in ICT (Trauth et al., 2008). Individual complexes of identities result in distinct experiences (Crenshaw, 1991; McCall, 2005; Shields, 2008; Bryant, 2017), and univariate approaches to “gender equality” are thus unlikely to achieve their intended impact apart from in highly specific circumstances (e.g.: Monroe et al. (2004) describe success in appointing women at elite US colleges established in the 19th century to educate the sisters of wealthy White men (p. 420-421)).

These calls for attention to intersectionality are not European, and thus less influential upon the HBP context. Furthermore, the role of women in ICT has received less scholarly attention in Europe recently (though see Walby et al., 2012; Pechtelidis et al., 2015). In a European context, “multiple inequalities” or “multiple discrimination” is the dominant frame within which identity intersections are addressed (Krizsan, 2012; Agustín and Siim, 2014). This is inadequate because it does not allow for compound or intersectional discrimination, precisely the phenomenon described by intersectional feminists and critical race theorists for decades. “Multiple inequalities” acknowledges that a single individual might be discriminated against in different situations for different reasons. However, different types of inequality are not structurally parallel or similar to one another (Verloo, 2006; Lombardo and Verloo, 2009); categories of identity do not have the same weight or influence in any situation; the model is narrow and excludes other approaches to inequality; and it omits the concept of class entirely (Kantola and Nousiainen, 2009).

Class or socioeconomic background is a significant factor in accessing career paths leading to a position in ICT or academia. Class and labour are considered in Marxist scholarship and feminist theorisations of gender in ICT (Fuchs, 2010, 2019; Adam et al., 2004). However, most approaches to diversity in ICT research (including intersectional works) lack deep engagement with class. The EPSRC Napier Report on Diversity mentions class in only a single instance, obliquely. This is concerning, particularly in light of the failure of the “multiple inequalities” framework to accommodate socioeconomic status and the innate, culturally contingent complexities in defining class.

There is another significant challenge to pursuing an intersectional agenda in European ICT (and therefore the HBP): Despite their prominence and centrality in intersectional scholarship, Black women have been “displaced from feminist dialogues about intersectionality in Europe” (Cho et al., 2013, p. 799). This is linked to current European attitudes toward the analytical utility of “race” or “ethnicity”, perceived as useful only in the USA and the United Kingdom (Cho et al., 2013; Lewis, 2013), which amounts to “an act of epistemological and social erasure—erasure both of contemporary realities of intersectional subjects […] […] and the history of racial categories […] […] across the whole of Europe” (Lewis, 2013, p. 887). Race and ethnicity, like gender and sex, are cultural constructs, and they play a major role in the exclusion of groups and individuals from participation (Rodriguez and Lehman, 2017).

Representation in the HBP

The HBP has an entire Subproject dedicated to Ethics and Society, which promotes the principles of Responsible Research and Innovation (RRI) within the HBP. Gender is an important part of RRI within the Horizon 2020 framework and is one of five thematic constitutive elements. Other approaches to RRI explicitly frame “diversity” as vital to truly responsible innovation: “Responsible innovation should not just welcome diversity; it should nurture it” (Stilgoe et al., 2013, p. 1573). Although the nature and utility of “diversity” is debated, there is agreement that it is not limited to sex or gender, with other categories of social membership more frequently associated with the term (Baker et al., 2016).

Beyond the overall under-representation of women in the HBP, they are most visible in managerial posts, and take on additional administrative roles. This is typical, and such trends benefit men’s research careers (Angervall et al., 2015). Of the 16 SP Managers in the SGA1 funding period HBP Blue Book, 11 were women. In comparison, there were 23 women listed as non-managerial leaders in the project Blue Book and 141 men. At that time, 7 of the 19 Ethics Rapporteurs were women and four of these were also SP Managers. HBP Subproject webpages present a similar narrative: of the research Subprojects, only SP12, Ethics and Society, had equal representation of men and women at the time of writing.

Policies to address gender in the HBP were designed to align with Horizon 2020 policy. Presented as “Gender Equality”, and outlined on the EC website:

In Horizon 2020 Gender is a cross-cutting issue and is mainstreamed in each of the different parts of the Work Programme, ensuring a more integrated approach to research and innovation. (http://ec.europa.eu/programmes/horizon2020/en/h2020-section/promoting-gender-equality-research-and-innovation)

A series of objectives, including: “gender balance in research teams”; “gender balance in decision-making” (with at least “40 per cent of the underrepresented sex” in panels and groups, 50 per cent in advisory groups); and “integrating the gender dimension in research and innovation”, are listed on the website and in the guidance (http://eige.europa.eu/sites/default/files/h2020-hi-guide-gender_en.pdf).

According to the website, achieving these objectives will:

• reach a critical mass of universities and research institutions in Europe which implement long-term institutional change through gender equality plans;

• increase the participation of women in research, improve their careers and achieve gender balance in decision making; and

• increase the scientific quality and societal relevance of produced knowledge, technologies and innovations by integrating an in-depth understanding of both genders’ needs, behaviours and attitudes. It also contributes to the production of goods and services better suited to potential markets.

The H2020 objectives appear to be aimed at increasing “diversity”, but this is only true if “diversity” consists of women already occupying intersectionally privileged positions (able, middle class, etc.). Furthermore, gender is presented as binary (“both genders”) and conflated with sex (“40 per cent of the underrepresented sex”). In the guidance document, this is only clarified with a brief reference to Londa Schiebinger’s work on sex and gender in innovation (Schiebinger, 2014). Race, ethnicity, disability and accessibility, class, religion, sexuality, age, gender reassignment (a protected characteristic in the UK) and related axes of oppression which clearly intersect with sex and gender do not merit mention.

The HBP has acted on these Gender Equality goals. An initial approach to addressing the underrepresentation of women and the issue of diversity in the HBP from January of 2017 (in the SGA1 funding period) included creating a Gender Advisory Committee and coordinating “gender equality activities” with support from consultants at EAF Berlin (Europäische Akademie für Frauen in Politik und Wirtschaft Berlin), a non-profit organisation dedicated to “diversity in leadership”. An HBP-wide Gender Survey was conducted from May to June of 2017, followed by the production of a Gender Action Plan in March of 2018, including targets and key performance indicators. The survey results and the Gender Action Plan were presented in a summary format in the Gender Equality and Diversity Conference.

Discussion

Conclusion

Despite the lack of access to the intended primary data sources, we have used first-hand observations of project documents and gender-related materials and events to demonstrate that the HBP approach to sex and gender up to March, 2018 is aligned with H2020 policy, and that several shortcomings are evident, which resonate with the concerns of intersectionality and labour scholars:

• A failure to acknowledge class, “race/ethnicity” and other factors outside of H2020 gender policy undermines attempts to recruit “women”.

• The resemblance to neoliberalised versions of feminism deployed to exploit women’s labour to the benefit of systems of oppression is uncanny.

• The focus on one gender may cause reticence to participate amongst women, and spur resistance amongst other genders.

• Targeting only “researchers”, “scientists” or “leaders” reinforces hierarchies and excludes many women from programmes.

• Likewise, a neglect of the areas in which women are overrepresented in the project leads to a narrowness of vision.

Whilst not intended to be comprehensive, the list below provides some ideas for developing related policy in large ICT projects:

• reflect upon the wording and vocabulary used when advertising funding, positions, etc.;

• include non-researchers in all aspects of work;

• include all genders in efforts to increase “diversity”, be clear that gender and sex are different and do not conflate gender with sexuality;

• fund childcare where children live;

• consider that individuals may have caring, household, or community duties apart from childcare, this may (a) incorporate intersectional positions such as income, age, religion, etc. and (b) be true of all genders;

• address education, project, meeting, and venue accessibility, take multiple religious calendars into account when scheduling meetings, and arrange social and networking events that are not centred around alcohol;

• make travel and hotel bookings through the project or institution so that reimbursement delays do not hinder applicants;

• when hiring, take breaks into account and recruit by career stage rather than the dates by which degrees or qualifications were obtained;

• arrange for long-term support of mentoring programmes and reward mentorship;

• make investments in educational ICT programmes for young children; and

• consider that high-quality research may not require researchers constantly moving institutions, and enhance support for international collaborations and consortia.

We have used an intersectional approach to open up the complex, culturally contingent space around gender and increasing diversity in ICT, but it is not clear what an intersectional HBP future would look like. Including sex, gender and other intersectional factors in scientific research increases research quality; more diverse representation in ICT is socially responsible and may stifle the perpetuation of intersectional marginalisation. But arbitrary requirements for an increase in “women” in the HBP are unlikely to have the impacts associated with diverse teams. Furthermore, would a truly intersectional HBP be 50 per cent “women”? Would the HBP serve as a representational microcosm of the population of the EU 28?

Any approach to diversity should be thoughtfully considered. The HBP has a unique opportunity to make a significant contribution to social good and equality within Europe and beyond, in addition to producing excellent science. Continuing, diachronic reflection on the impact of intersectional efforts in the HBP (and ICT more broadly) is a potentially productive line of future research.