User-Centred Design is not so user-centered.
Biases can permeate our designs. What should we do about it?
As I have argued before, the relationship between designer and user is political. However, it is rarely recognised as such, and the responsibility of designers not appropriately acknowledged.
The critique on this omission finds an exemplary expression for designers working in collectives of similar demographic and ethnographic background, as exemplified in this photo.
More popularly, this is pointed out by the present supremacy of cis-gendered white males in Silicon Valley, as reported on by data reporter Sinduja Rangarajan. They are widely criticised as their ‘universalizing aspiration about design and use of information and communication technologies’¹ represents only one worldview, their own, through what Google’s former Vice President of Communications Jessica Powell argues is a ‘monoculture of thought’². Similarly, designer Lu Han claims that ultimately, such uniform collectives are not increasing the understanding of their design’s impact on a system sufficiently, as they fail to develop sensitivity for the experience of other groups.
This issue is worsened when — as designers — we apply User-Centered Design methods to establish representations of anticipated use and user. Even when we empathetically engage and collaborate with users with good intention, we are unsuccessful in eliminating our individual biases from the design process. Furthermore, we hold a privileged role, acting as a ‘gatekeeper’³ for information collected about users as well as for the decision-making process.
Such a position makes the research on the user subject to bias and assumption. Industrial design engineer Marc Steen claims that we will filter what we see and hear from users and stakeholders by what we view as significant, ultimately making it difficult for them to change our initial biases and assumptions.⁴ Sociologist and economist Tim Seitz suggests, that under the pretence of understanding and reflecting the real needs of a person, User-Centered Design practitioners, in fact, define the user.⁵ With that, we ultimately construct a specific reality to fit in our ideas for a product or service.⁵ Similarly, Professor of Urban Planning and Public Service Natasha Iskander argues that the choice and characterisation of the user and their needs will implicitly be refractured through the designer’s personal experience and priorities, in order to transform the collected information’s ambiguity into guiding information. Ultimately, those will shape the design.
Researchers in the humanities are aware of biases in academic research throughout ‘planning, data collection, analysis, and publication’⁶ and have developed methods attempting to identify and avoid them. However, User-Centered Design research is yet to establish adequate mechanisms. Iskander suggests that subjectivity in research is generally inevitable, which is why academic disciplines relying on empathetic engagement for data collection put emphasis on the importance of the researcher’s identity and political position. Nonetheless, we are rarely classified in such manner. With that, we can unwittingly reinforce ‘bias and inequality’ as well as ‘stereotypes and structural oppression’.
Biases are ubiquitous in everyday software and its interfaces, such as gendered friend or group icons on Facebook, subordinating female representations, but also depicting predominately white ethnicity. An infamous example is Google’s autocomplete suggestions that reflect female oppression by completing terms as ‘women should’ with sentences like ‘stay at home’, ‘be slaves’, ‘be in the kitchen’ and ‘not speak in church’. Although both corporations reacted to partially eliminate both instances, bias seems to remain inherent to the User-Centered Design process, as it can still be found in emerging technology such as voice assistants. Research on the user’s ‘ideal AI’ results in the preference of a human-like personality with a female voice when given male and robotic as voice alternatives.⁷ Likely to be based on similar research, the default voices of Amazon’s Alexa, Google Assistant and Apple’s Siri are female, revealing gender role stereotypes permeating the design. Even though the providers allow users to change the voice of their assistants to male, the design still represents a worldview of binary gender. A recent interdisciplinary project designed a ‘genderless voice’ to be adopted by voice assistants, as:
‘Technology companies often choose to gender technology believing it will make people more comfortable adopting it. Unfortunately, this reinforces a binary perception of gender, and perpetuates stereotypes that many have fought hard to progress.’ ⁸
Such examples show how design can reproduce specific ideologies. User-Centered Design practitioners are attempting to solve this issue by involving more diverse sets of stakeholders and potential users throughout every part of the process. However, critics argue that even though engagement with the user is aiming for openness, to learn from users, it is yet moving to closure by choosing them and drawing conclusions about them. Steen suspects that, while doing so, we are not always aware of the effects our backgrounds and methods have on this decision-making process.⁴ This suggests, that by applying User-Centered Design methods we are not actually challenging our assumptions but are rather looking for their confirmation.
However, we should be aware of this issue and the responsibilities we carry beyond a product, service or experience. The relevance and importance of this critique is heightened by the evolving role of the ‘designer’, and our growing involvement in large-scale social challenges like public services and politics. Therefore, due to the scale and speed in which it continues to be uncritically applied, User-Centered Design can have adverse consequences. Thus, we are required to revisit our methods and focus, not only on the enlarging design space but on our increasingly holistic responsibility.
Earlier, I have presented the intrinsic value and underlying accessibility of User-Centered Design for a broad base of practitioners as well as its potential for democratic empowerment. I, therefore, do not conclude in the claim to abandon User-Centered Design. Instead, I urge an evolution of the current approach to take on a more holistic and responsible outlook that moves beyond commercial return and forward towards sustainable and progressive design. This evolution ought to include the awareness of the our political responsibility for users. Furthermore, we should understand and explore our own biases and prevent them from entering the different stages of the User-Centered Design process.
Read more about the History and Nature of User-Centred Design and about its political potential.
¹ Bardzell, S., 2010. Feminist HCI: Taking Stock and Outlining an Agenda for Design. Atlanta, CHI 2010: HCI For All, pp. 1301-1310.
² Powell, J., 2018. Silicon Valley’s Keystone Problem: ‘A Monoculture of Thought’ [Interview] (2 October 2018).
³ Jones, P. H., 2014. Systemic Design Principles for Complex Social Systems. In: G. Metcalf, ed. Social Systems and Design. Translational Systems Sciences. Tokyo: Springer, pp. 91–128.
⁴ Steen, M., 2012. Human-Centered Design as a Fragile Encounter. Design Issues, 28(1), pp. 72-80.
⁵ Seitz, T., 2017. Design Thinking und der neue Geist des Kapitalismus. 1st ed. Bielefeld: transcript.
⁶ Pannucci, C. J. & Wilkins, E. G., 2010. Identifying and Avoiding Bias in Research. Plastic and Reconstructive Surgery, 126(2), pp. 619-625.
⁷ SYZYGY, 2017. Sex, lies and A.I., New York: SYZYGY.
⁸ Copenhagen Pride, Virtue, Equal AI, Koalition Interactive & thirtysoundsgood, 2019. Meet Q The First Genderless Voice.
