On December 1st 2024 an interdisciplinary group of researchers gathered at UNSW Sydney on unceded Bidjigal land, gathered around the question ‘what can’t data hear?’ The day involved presentations, provocations, discussion and reflection that explored the implications of the datafication of the social world for processes and practices of political voice and listening. Throughout the day, participants considered questions around the increasing ubiquity of datafied voice and listening in a context of widespread neoliberalism. This blog post overviews some of the discussion generated through the workshop. The full program, including abstracts and bios from each speaker, is available here. The workshop was convened by Associate Professor Tanja Dreher, Dr Poppy de Souza and Data Stories post-doc, Dr Danielle Hynes with excellent and essential assistance from Mitchell Price.
Questions of what is left absent when voice and listening are increasingly datafied within neoliberalism are relevant across many topics and disciplines, including (to name just a few) health, housing, education and politics. Reflecting this, the workshop was split into two broad themes: housing and care. The workshop was split into 5 sessions; 4 involved presentations from speakers followed by short discussion, and the final hour was dedicated to discussion and reflection between the whole group. The names of each presenter and a brief word on the focus of their presentation is included here (check out the program for full abstracts of each presentation and a little more detail on the premise of the day).
Housing
- Claire Daniel – Platformisation of urban planning
- Sophia Maalsen – Know your Landlord: Bringing attention to data silences
- Linda Przhedetsky – Being ‘known’ and ‘unknown’ by algorithmic systems in rental housing
- Alistair Sisson – On counting homelessness: The ambiguous function of By-Name Lists
- Chris Martin – Rental vulnerability, advocacy and assistance
- Jacek Anderst – Social housing waitlist data invisibilising housing needs of Aboriginal and Torres Strait Islander people
Care
- Georgia van Toorn – Ontological misrecognition: Datafication of support plans in the NDIS undermining self-determination
- Andrew Whelan – Big mental health data for care-less futures
- Linda Steele – Disability, violence and reparative justice
- Poppy de Souza – (Un)training the algorithm in an age of datafied (non)care
- Mitchell Price – Speaking against oneself
- Diane Stapleton – Desire at the heart of datafication: trans visibility and trans erasure
The remainder of this post recounts some of the discussion that was generated throughout the day.
Image: Workshop participants in discussion
The day began with a focus on housing, generating questions around the tension between platform values and public values, and the increasing dependency of public institutions on platforms in a context where public institutions are increasingly reliant on, and providing funding to, private platforms to facilitate urban planning. The trend of power imbalances being exacerbated through these platforms is not unique to planning and housing, and participants noted there may be a possibility of alliance building across different areas in relation to this.
Discussion following the second session, also focused on housing, generated questions around lived experience/lived expertise, and how we might better attune to and present lived expertise as a vitally important way of understanding how data driven systems works in practice and impact people. Participants noted that it is important to begin with strength based, rather than a deficit based perspective – understanding that lived expertise can constitute forms of knowledge that contribute to theorising and conceptualising how things work, not only descriptive story, or trauma focused (as important as these contributions also are). People impacted by datafied systems use innovative methods of ground truthing to find ways to function/survive within systems that influence their lives, such as (in a social housing related example) driving around town to figure out which properties are empty and requesting to be housed in these specific properties, challenging claims that ‘nothing is available’. The session concluded with a caution; lived expertise can be incorporated into larger systems without meaningfully affecting change, potentially co-opting the potential of this type of knowledge. This theme was to recur throughout the day.
Moving to a focus on care (and acknowledging that care and housing are deeply intertwined), conversation revolved around the importance of mutual recognition, the social nature of identity formation and the difficulty (or impossibility) of fulfilling the need for human recognition of experience and identity through datafied systems. This surfaced across conversations about the datafication of multiple aspects of healthcare, including support plans in Australia’s National Disability Insurance Scheme (NDIS), mental health, and violence and disability. Also prevalent in this conversation was the individualisation of proposed datafied solutions to what are often social problems. Certain values come to be baked into these systems, and these values can be obfuscated. It is important to interrogate how these values come to exist within these systems, and how they are secured.
Important in this discussion was challenging the view that all that is needed is more/better data, and bringing in the idea of reparations. For example, NDIS support plans, plans that are specifically targeted to individuals which are intended to name their self-defined goals and the supports needed to achieve those goals. To be effective, these plans must be monitored and updated continually. An algorithmic system is used to generate a blueprint of support plans for individuals, and the combination of algorithmic bias and funding cuts means that what is meant to be a basic blueprint reviewed carefully by an expert, can become the final support plan without sufficient review. Linked to this was bringing into the conversation the long history of diagnostic criteria that is grounded in eugenics, and the close ties this has to colonialism – noting that bringing the concept of disability into Australia has been part of the violence of the settler colonial project.
Following this the conversation moved to another aspect of care, that people do care about stuff; and the questions were raised ‘how can we recognise and honour what people care about, and recognise the systems that corrode care and make it difficult? Is there another way we could organise our societies?’ One point to follow in response to this is valuing the interdependent and collective nature of care, linking to the history of disability justice movements. Within this exists a tension – at times people are able to build collective, peer led mutual aid groups that provide mutual support, but the existence of such groups may be taken as an opportunity to gut the state aspect of support provision.
Image: Mitchell Price and Associate Prof Niamh Stephenson
Finally, we concluded the workshop with a reflective session. Throughout the day everyone had access to blank index cards, bringing an analogue element into play as we wrote thoughts, questions and points towards a manifesto on these cards. These served as prompt for our final discussion, and were collected and transcribed by Poppy de Souza (thanks Poppy! More to come from this soon). What is written here arose from the verbal discussion, rather than the index cards. Discussion focused on a few themes: the politics of (in)visibility and transparency, how to interrogate the values embedded in systems, what constitutes ‘enough’ data, and the importance of attending to history.
There are tensions around the value and politics of visibility and transparency. Thinking about theories of change and transparency; while transparency is an important principle, it isn’t necessarily clear how it can activate change (as some of the workshop participants have written about). It’s important to remember that transparency and visibility are not end points in themselves – a transparent system may well be unjust and uncaring. Related to this, it is vital to look at the values embedded in systems, more transparency without shifting embedded values will not activate change. A participant asked, ‘how is it possible to understand what values are embedded in systems?’ Some methods suggested included:
- Discourse analysis and critical discourse analysis
- Looking at resource allocation/follow the money, e.g. the Australian Tax Office is not sufficiently resourced to investigate/recover corporate tax fraud, but it is highly resourced when it comes to investigating welfare fraud, what value does that represent?
- Exploring ‘at what point does it stop being possible to see?’: looking up, tracing a system, where is the point that we can’t see anymore. That is a value project, when does it become possible to obfuscate (e.g. resource allocation, values, who’s involved)
- Looking to perverse outcomes, such as people seeking a type of social housing that doesn’t meet their needs because they know the wait time is shorter for that housing type and that in turn being interpreted as a reflection of the demand for that type of housing, or homelessness By-Name List numbers going down being celebrated as representing a reduction in homelessness when it really represents disenchantment with the system from service providers and people experiencing homelessness and a lack of resourcing for staff to add people to the list. Perverse outcomes such as these show cracks, that the system can never fully recuperate what it claims to represent.
At what point do we have ‘enough’ data? Particularly when it comes to cases where people have shared their stories, their trauma, and there has been little or no action. In these cases the call cannot be ‘we need more data’. The data needs to do something/something needs to be done with it.
Finally, and fittingly, came the reminder about the importance of attending to history, and always asking ‘are the problems and issues with data and datafication new, or are they extensions of existing patterns, systems and power relations? What precisely is new or different?’ Categorisation has long been an essential element of colonisation – the process of dividing and naming. Digitalisation and datafication are also about splitting and demarcating. The digital at its heart is a system of 0s and 1s, a binary system, and there are certainly things that the binary can’t hear.
Some key overlapping themes throughout the day were:
- The importance of hearing/listening to lived experience/lived expertise as data
- Attending to perverse outcomes of data generation/collection/analysis
- Why collect data and what is ‘enough’ data? This linked to theories of change (what is the intended outcome of collecting this data? How will it change things?), data saturation (is collecting more data about this necessary?), and data collection as procrastination (rather than doing something about an issues, are we instead collecting more data about it? Is more information truly needed, or is action needed)
We would like to extend an enormous thank you to all the presenters, facilitators and participants for their generous and generative contributions throughout the day.
