Jutta Treviranus is the Director of the Inclusive Design Research Centre (IDRC) and professor at OCAD University in Toronto, formerly the Adaptive Technology Resource Centre. It was fascinating to think about the links that the DDI might have with the Centre, and Jutta’s illuminating and provocative talk challenged many assumptions and provided many opportunities to consider where we can incorporate some of the ideas she talked about. She wanted to talk to us about challenges she is currently struggling with, and discuss ways forward, hopefully gaining some perspective on some fairly “gnarly” and complex problems. She wants to inspire us to think about the mechanisms that are used by governments, and the systemic processes instilled in institutions, and how digital systems and AI play parts in them.
For Jutta and her team, the biggest challenges are being accountable to, and for, the outliers and small minorities in a democracy that are often marginalized, and either negatively impacted by decisions which are made for the majority, or just not considered at all. Her provocation to the DDI is to consider how we go beyond majority rules? To take account of those who are not part of the majority, but often have very critical needs. Her driving impetus for all her work is the premise that “diversity is our greatest asset, inclusion is our greatest challenge,” which she admitted may come across as a trite soundbite, but at its heart is the kernel of her purpose. Currently, neither is served democratically by the way that we make decisions.
Counter to what you might expect, this has been particularly prevalent during covid, as a sort of social Darwinism has been created, as ensuring the survival of the fittest seems to have been the prevailing decision maker. Jutta called for a reconsideration of Darwinism as a result, in that his true theory of evolution needs to be wholly understood in seeing diversification as necessary for evolution as it that which leads to advances. We need vulnerability and fragility to survive. It compels us to create a system that will work for us when we are vulnerable – covid has proven that. As she so neatly put it, none of us are safe until we are all safe.
Since 1980, Jutta has established a very large dataset of human requirements with regard to UX and digital systems, which she has recently been able to plot in a 3-dimensional multivariate cluster graph that she describes as a human starburst shown in the image below. Dots in the middle, that represent the norm or average, are close together and represent people who have similar needs, dots toward the edge, that represent the data tails or outliers are far apart, meaning the needs are dissimilar. The majority of solutions, for advertising, health, everything really, are all focused on the people in the middle, and the people on the edge are left out as either being too niche, too costly etc. It is these outliers that Jutta and her team focus on.
The impacts of the ‘best-fit’ scenarios ripple through every aspect of society; design that fits the 80% is cheaper in the greater production scheme, and therefore it is harder to get accessibility for those 20% who don’t fit as shown in this fantastic hierarchy of needs below. The 80/20 Pareto principle that was mischaracterized by Robert Koch has done a lot of damage. If we attend to the 20% in policies and design, the remaining 80% has room for growth and change. This is where there is potential for greater inclusivity.
Data is, of course, one of the biggest exponents of the problem. Social justice strategies are based on data, on bounded identity groups, categorization. What happens if you fall through the cracks or are not included? The understanding of equity programs by politicians therefore miss the most vulnerable, and then it becomes those individuals’ problem for not fitting. Pareto called the 20% the “vital few,” which are often today simply been called difficult. Jutta described the “cobra effects” of unintended consequences of over-simplistic ‘solutions’ to complex problems. As an example in the academy, the recent cutting of humanities at Laurentian University is a representation of what is considered important by those who make decisions. Linear thinking exacerbates this, so it is a call to think creatively that will inspire truly holistic strategies. There are also many examples of cobras in the Government too, particularly where equity is confused with equality. What is happening currently is that we are “optimizing toward homogeneity and conformity,” instead of moving towards diversity.
When thinking about the impacts of innovation, according to Thomas Friedman, technology is adapting exponentially, but humans are only adapting linearly, as the below graph shows. This actively impedes adaptability, and using stock phrases like “people like you like this,” ‘smart compose’ and technologies like this are reducing the vocabulary, mechanisms and processes to the average. Jutta also proffered that connected technology is impeding human connectivity by being device mediated, thereby creating virtual tribes, which results in an incomplete representation of self, with no behavioural repercussions. The societal outcomes for which are very disconcerting.
In cementing this unsettling prospect, Jutta described how recent experience with Bill C7 demonstrates this trend. The bill about medical assistance for helping the terminally ill to die comfortably recently passed, and involved a highly politicized and emotional set of decisions, in particular where it came to the sections about which disabled medical conditions were included. By being able to choose this as an option for disabled people could lead to devaluing of lives that they would otherwise chose to continue, and this choice of “death over disability” could feed into a form of eugenics. Bill C7 did pass as it stood, without changes despite the concerns for disabled people. The reason it wasn’t stalled was due to data. There apparently was insufficient data or evidence that any further debate was needed, evidence to the contrary being dismissed as anecdotal versus the 300,000 average Canadians that were surveyed who provided quantified dataset that Canadians supported the amendment. As Jutta put it, this resulted in a “reduced truth” being presented, and that the impact was only seen to be fair on the majority not on the minority.
AI therefore presents these sorts of decisions as much more dangerous, without adequate representation of all interested parties, equality being considered more important over sensible equity of consultation with affected parties. This leads to the rationale for decisions not being considered, enforcing black box thinking that only reinforces patterns of the past. A true and practical ethics regulation for AI is one significant option for having more of an impact on diversity considerations. One other scenario that she didn’t have time to get into fully is how anomalous and unexpected people (such as someone propelling themselves backwards in a wheelchair) are treated in machine learning models of self-driving cars in intersections.
In closing Jutta reminded us that these power tools of AI are not currently set up for diversity; perceptions of privacy are useless for highly unique people, and informed versus implied consent issues are completely inaccessible for those demographics on the edges. As a result, Jutta’s call us to include the diverse margins! This will help to reach a more dynamic equilibrium with increased diversity. If you want to read more, you can read Jutta’s Medium article on inclusive design here, or contact her directly here.