Two children on steps with a clipboard trying to figure something out together.

The paper falls far short in its appreciation for human-centred design, and this is quite typical of contemporary approaches to digital identity.

Originally posted to the generative identity website.


The OpenID Foundation has invited comments on a draft paper titled Human-Centric Digital Identity: a Primer for Government Officials (July 7, 2023), referred to from here on simply as 'the paper'.

It is critical to understand how the paper's authors understand human-centricity, so I begin there. I contrast this with a necessarily broader and deeper scoping of the human-centred design approach; when human is confused for individual, as here, human-centric concerns can only be addressed by de-centering the individual.

I argue the imperative to move beyond the human-centric paradigm if only to best serve what most people will think of as human-centric; at least with the future's benefit of hindsight.

I conclude that this paper does not achieve its stated purpose of priming government officials sufficiently or accurately. I offer an explanation relating to a non-human-centric standpoint shared by too many members of the OpenID community and too many actors in government and civil service. As this standpoint has informed the design of digital identity for decades, it's very difficult to retrofit a contradictory design approach.

Where I have had the choice to write a thesis or offer something people might actually read, I've opted for the latter. I point this out for any academics ready and eager to critique my critique!

The paper's regard for human-centred design

The paper attributes (Nyst et al 2016) with pre-supposing "human-centric objectives" (line 298). The reference in question doesn't actually refer to human-centric at all but rather uses individual-, user-, person-, and consumer-centric, without distinguishing one from another. The reference describes associated qualities:

  • Entailing "greater control over their data" through the provision of personal data stores
  • Greater control over access to attribute data
  • Privacy preserving approaches
  • Encouraging the system to be "focused around the individual"
  • "Making the individual the control point for their digital identity".

The paper's main section describing human-centred design comes later (lines 481-594). Reference is made to (de Brisis and Brennan, 2022, Universal Digital Identity Policy Principles to Maximize Benefits for People: A shared European and Canadian Perspective), specifically that its human-centric principles epitomise those of a variety of other sources. This reference uses human-, people- and user-centred, again without distinction. The aspiration is framed in terms of ensuring that "[p]eople's needs and rights are prioritized", with five assertions:

  • Digital identity must be inclusive
  • Digital identity adoption must be voluntary
  • Digital identity must be resilient
  • Digital identity policy must be intuitive for both institutions and people
  • Privacy must be central to digital identity policy.

Reference is made to (Trust Over IP Foundation, 2022, Overcoming Human Harm Challenges in Digital Identity Ecosystems), a draft report that was published as version 1.1 in May 2023. This is an odd reference to make here. The ToIP report only mentions people- and human-centric in passing. Rather than a human-centric approach per se, it focuses on claims of following a socio-technical systems (STS) approach. As I argue in (Sheldrake, 2023, Live or DIE), the ToIP report does not in fact evidence the claims but rather conveys significant shortcomings that can only lead the reader to reach the opposite conclusion.

The OECD reference (2023, Recommendation of the Council on the Governance of Digital Identity) sticks with "user-centred" where user "refers to a natural person or a legal person, or to a natural person representing a natural or legal person." It asserts that developing user-centred and inclusive digital identity entails responding to the needs of users and service providers, and minimising barriers to entry. It all feels very perfunctory, as one might expect when user is preferred to human.

While reference is made to (unicef, Human Centered Design 4 Health), the reference is too brief. It picks up solely on the need for engagement and iterative design but leaves the rest.

There's no arguing the paper's inclusion of inclusivity under the heading of human-centred design, ditto value-sensitive design (VSD). But again, the paper is too selective in bringing VSD into the process. The selectivity is not noted let alone explained.

I speak to both of these selections / omissions below.

Human-centred design (HCD) today

Umuntu ngumuntu ngabantu.

   Zulu: A person is a person through other people.

Ndiri nekuti tiri.

   Shona: I am because we are.

HCD today — by which I mean practiced as well as we understand it — incorporates such relationality, such complexes, as encapsulated by these African philosophical phrases.

Within dynamic relationships, within our communal complexes, we find the fundamental process of meaning-making. It's what we do together. It's what we do together better than any other species on Earth. In fact it's essential to what makes us human, so it really doesn't get much more human-centric than this!

My use of complexes is deliberate. Two potatoes are two potatoes. Ten thousand bricks are ten thousand bricks. Potatoes and bricks don't exhibit emergent properties as they come together. People do. The meaning of human in human-centred must be read therefore as being human in relationships, in communities, and most definitely not individual or user or consumer.

Invoking indigenous wisdom doesn't always succeed in conversation within the contexts of the conventional western mindset, so I will also pull from one of the paper's references. While the paper notes (line 499) that "human-centred design processes involve broad engagement and iterative design test sprints that build empathy for those who will rely upon the system", citing (unicef, Human Centered Design 4 Health), it makes no mention of one of the key messages on the reference's homepage, i.e. the unavoidable need to take "a system’s view of an ecosystem affecting people’s behaviour." The referenced website never talks in terms of the individual, not once, but rather people, groups, families, and communities.

To take a human-centred design approach is to necessarily emphasise dyads, family units, local communities, and society as much as the individual, and to reflect deeply on relationships and learning and "the ultimate meanings which the product, system or service will either occupy or create within the psychological, sociological and societal space of the individual." (Giacomin 2014)

Human-centred design is concerned less with assuring that artifacts work as intended (by their producers, designers, or other cultural authorities) than with enabling many individual or cultural conceptions to unfold into uninterrupted interfaces with technology. (Krippendorff 2004)

In truly stark contrast, the paper talks in terms of the individual throughout, most notably in the very first sentence (line 2) and in fact in its very definition of identity. To be fair to the authors, this error is endemic.

And so to understanding human identity …

Conceptualisations of identity across all variety of the social sciences consider identity to be a capacity for and never-ending process of sense- and meaning-making. I find this surprises many technologists working in digital identity. This understanding produces a fascinating and vital reflexivity (a social science term referring to circular relationships between causes and effects).

Human-centred-design actively supports meaning-making. Human identity encompasses a continuous process of meaning-making. And so human-centred-design-for-digital-identity may be re-stated as meaning-making for meaning-making, a continuous and ideally virtuous loop.

As this is a forever dynamic process, any technical device that might disallow, resist, or attenuate contextual meaning-making and so change — generativity — is inevitably being misapplied here. How do the celebrated qualities of the cryptographic fundamentals of digital identity feel to you in this human-centric light?

Value-sensitive design (VSD)

"This paper recommends an inclusive, value-sensitive approach to Human-Centered Design that is underpinned by standards." (line 495)

"Pillar #1 Recommendation #2: Deliver a Values-Sensitive design process that reflects the risks and benefits to stakeholder groups." (line 803)

According to the authoritative website on the topic (vsdesign.org), VSD engages "human values in the design of tools and technology to support human flourishing". A value is "what is important to people in their lives, with a focus on ethics and morality." It considers "individual, group, and societal levels of analysis" and observes that "in the complexity of human relations, values sit in a delicate balance with each other. This framing positions designers and researchers to emphasize moral and ethical values, but to do so within the complexity of social life, and with recognition for how culture and context implicate people’s understanding and experience of benefits as well as harms and injustice."

The paper's section on VSD (lines 544-567) presents some advantages of adopting VSD to the reader. Here we find the sole reference to groups that isn't preceded by 'working' or 'stakeholder', although it remains too strong a possibility that the authors consider this shorthand for stakeholder groups, else a group considered as merely the sum of its parts. Either way, the true power of VSD is kept at bay for some reason.

The paper makes no reference for example to human flourishing or ethics or morality. Nor does it point out that values vary from community to community and from time to time whereas current digital identity schema offer a one-size-fits-all with some fearsome path dependencies. (Variations within a schema should not be confused for variations of schema.)

Beyond human-centred design

As valuable as human-centred design undoubtedly is, it's most often stretched to the point of inadequacy in the contexts of complex adaptive systems and, as (Jones 2022) notes, it is now being combined with systems thinking, futures thinking, and change theory to set up the conditions for learning experiences that are responsive to such design challenges. We are undoubtedly in such contexts here.

"Systems thinking is 'contextual,' which is the opposite of analytical thinking. Analysis means taking something apart in order to understand it; systems thinking means putting it into the context of a larger whole. … [It] means a shift in perception from material objects and structures to the nonmaterial processes and patterns of organization that represent the very essence of life." (Capra & Luisi 2014)

Systems thinking is integral to socio-technical systems (STS) approaches. STS considers the intermixing of technical and social systems with the diligent application of the respective disciplinary knowledge, and ideally the development and application of transdisciplinary knowledge. Both (large scale) technical systems and social systems demand systems thinking, and so then does their inseparable interweave.

In (Sheldrake, 2022, Human identity: the number one challenge in computer science), I argue that the technologization of human identity is a complex of embedded, embodied, extended, distributed and collective cognition (Smart et al 2017), and when life is considered a trilogy of biological structure (autopoetic unit), environment, and cognition (Capra & Luisi 2014), the technologization of human identity is in fact a living process. We should design for psychological, sociological, and ecological health — self, culture, and nature — rather than mere process efficiency or convenience.

I've found a similar conclusion presented by Fjord (2020), a design and innovation consultancy owned by Accenture, albeit applying more expansively. Inspired by a conversation (2018) between Gerry Scullion and John Thackara, itself riffing off Thackara's 2005 book In The Bubble: Designing in a Complex World, Fjord writes:

"... people’s focus of desirability, viability and feasibility is moving from 'me' toward 'we'. As this accelerates, user-centered design will feel increasingly selfish … [We can] see an evolution in design from user-centred to human-centred and now life-centred design ... starting to edge away from designing for one to designing for collective — i.e. the entire planet.”

The following diagram (Borthwick et al 2022) portrays life-centred design principles adopted from (Gibson 2001) and supporting frameworks and actionable methods.

Listing 7 principles: integrity; sufficiency & opportunity; equity; efficiency; democracy & civility; precaution; immediate & long-term integration

I can only think of life-centred design approaches to the digital mediation and augmentation of human identity. As I have written previously, we must approach the technological augmentation of human identity as a living process and its domain an ecology.

Conclusion

The paper's description of human-centred design is insufficient and inaccurate for its omissions and for its dedication to the individual object rather than being human. It might be said to invoke those qualities that could throw a warm light on the OpenID community's work while leaving out the qualities that might leave the reader concluding otherwise.

Its nod to value-sensitive design is guilty in the same way, but perhaps all the more so given, well, the values underpinning VSD.

It is imperative that we find a way to work together to remedy the paper after this request for comments, but as should be evident from this comment right here, there is no shortcut.

There is no shortcut for the simple reason that the cause is paradigmatic. Computer scientists, information technologists, and those serving in our governing bureaucracies share one paradigm of identity that I refer to as noun-like. The social sciences and the humanities acknowledge the development of such a paradigm, but recognise that it's not ontologically equivalent to the verb-like paradigm. The latter attempts to describe how life is, how we humans understand each other and ourselves and the world around us, whereas the former attempts to find a way to 'run things', so to speak. Governments. Large organisations. Bureaucracies. i.e. the kind of things that are rarely if ever described in terms of being human; often quite the opposite.

When our lives only infrequently bumped into the 'running of things' this may have been fine, but the technological manifestation of the bureaucratic noun-like paradigm is being set up to pollute the information ecology of human nature and human culture, as I explain in some detail in (Sheldrake, 2022, Human identity: the number one challenge in computer science). This is anti-human-centric.

The authors of this paper risk appearing to engage in some serious 'HCD-washing' and 'VSD-washing' in a remarkable parallel to the 'STS-washing' evident in the recent ToIP report on the human harms of digital identity. I'm putting that as politely as possible. We must all be more blunt if this carries through to the final version.

"... human-centred design is increasingly seen and applied to achieve short-term gains for businesses and investors. This economic focus has diminished the potential of design to serve as a tool that enables deep exploration of decisions before they are made (Dunne & Raby 2013) to ensure that they are the right decisions. Focusing on short-term gains has also drawn attention away from genuine global problems, future generations and the health of the planet — in many cases, human-centred design has even contributed to these problems." (Borthwick et al 2022)


Update 27 September 2023

The OpenID Foundation published the final version of the paper 25 September 2023. The very poor but lamentably expected news is that nothing changed materially in the context of my critique here.

This critique and my subsequent conversation with one of the paper’s lead authors last month did have some small influence. The final paper now notes (p35) a distinction between human-centred and merely person- or user-centred, and incorporates part of the (Krippendorff 2004) quote. But it’s just a fleeting acknowledgement rather than anything that might be said to have been recognised anywhere else in the paper. It certainly appears to have had no impact on the contributors’ worldview, which should be nothing less than deeply concerning for the rest of us.

###