Inclusive Data Research Skills for Arts and Humanities/Deconstructing data methods and decolonising approaches

Defining the challenge, mapping out contexts and key stakeholders

When defining challenges, it is necessary to deconstruct and unpick the kinds of thinking, epistemologies and colonial practices that come with data tools, skills and methods. Embracing the collaborative and plural nature of a hackathon, we contextualised the relationship between coloniality and data methods through multiple perspectives and disciplines. We began to map out who the key stakeholders are, what are the core systems, institutions, frameworks we need to address.

1. Defining our key terms:

data

data methods

data epistemologies

coloniality

decoloniality

2. Reflections:

How do we encounter coloniality or de-coloniality from our lived experiences and/or within our particular disciplines? Where do data oriented computer tools or data practices come into this? What is the data being collected, who or what is being observed and who is observing, how is the data being labelled or categorised, what is the data used for?

How are data methods embedded within colonial disciplinary priorities and biases?

For instance: categorization, thinking in binaries, quantification, reduction, cultural privileging of “rational” science, the myth of ‘objective data’, the capitalisation of data, extractive data practices, bias, environmental extraction required for data processes to run.

Some of the areas discussed include:

• Conversational generative AI is mostly trained on a western, capitalist, imperialist, patriarchal corpus and will drown out other voices
• English mono-language based research
• The prioritisation of written, digital, 'big' quantitative data
• The distinction between right and wrong. In terms of education, decolonial methods in new schools are frowned upon or unsupported by the government a lot but helps students.
• Difficulties putting theory into practice
• Retroactively filling historical / research gaps. Revisionist thinking back and forward

How do the limitations of data or computational tools force certain worldviews or certain ways of thinking? Include examples of systems participants use within their field.

E.g. AI, search engine systems or research databases, social media analysis tools, facial recognition systems, surveys and questionnaires offering limited identity categories

• How can we value, recognise, perpetuate and build upon data from non-written groups such as indigenous and first nation peoples?
• "On the Dangers of Stochastic Parrots: Can Language Models Be Too Big?" - by Bender et al, 2021
• If it can't be connected, it doesn't matter
• The area (geographic), assumptions, healthcare services, neurodivergent people, education system in the UK which is a settlement country, often majority of education is only in english.
• online, computer and technological sources
• Geographic Information Systems (GIS) have no room for qualitative data
• How can we integrate indigenous methodology and alternative methodologies into canonical methods taking a BOTH, AND approach?

Who are they key stakeholders / players involved in colonial data methodologies? Who and what are we needing to address when we approach these challenges from a decolonial lens?

• Data method researchers, The domain knowledge experts, Data recipients
• Sponsors, curators and participants (data gatekeeping), transparency, supportive of people, collaboration, building trust, reciprocation.
• Myself, UKRI, renewed use of hidden discriminatory parameters
• acknowledging stakeholders "sociological imagination" and their validity as researchers. THE ETHICS OF STORYTELLING
• managerial data people, knowledge hoarders, 'insight' commercial companies and more
• people in positions of power to influence such methods

DATA GAPS / ABSENCE

E.g. research gaps, theoretical gaps, access to computation and data. Do we need to fill gaps or how can they be useful? HI-RES/LO-RES BIAS certain geographical locations have concentrated data and concentrated technologies

Some points of discussion included:

• Participation of marginalised groups as practitioners and theorists
• How can we teach, educate and create a decolonial framework in a society that is largely colonialist?
• Invisible Women by Caroline Griado Perez really highlights the manifestation of such gaps
• Revisiting methods and theories that are currently in use
• gaps in data lead to a half-told truth. More inclusivity and plurality in data collection, storing and analysing will lead to filling gaps.
• Parachuting in design. Be more available in terms of interaction, don't judge, be a facilitator not an expert
• Creatively fill and show diversity in life experiences

Alternative Lens, mapping out key approaches and movements

What does a decolonial lens look like when working with data? What can we learn through examining the broad approaches different groups are taking in order to counter hegemonic powers?

Themes and approaches include:

Participatory Action Research / Co-design

• Data Walks Alison Powell [1]
• Ethnography
• Citizen Science eg. IMPETUS [2]
• Role playing and games
• Equity film making [3]
• Any method that redresses or aims to redress the imbalance between powerful researcher extracting data from data subjects
• Action research [4]
• Using your own body as a site of research, self-realisation
• Utopia
• Design imaginaries

Democratising access to design tools and skills

• Decolonizing Reader (resources)_Collaborative (open) edition [5]
• Github, open source, open access, open data
• Low tech approaches - drawing data
• Creative non-tech approaches to data literacy through the arts and community eg Data Murals [6]
• MediaFutures Toolkit - focusing on AI/art interactions but with lots of lovely open access data tool resources [7]

Pluralist lenses

• Indigenous voice
• TEK (traditional ecological knowledge)

Using data/ creating datasets for justice

• Atlas of Environmental Justice - documenting global environmental conflicts [8]
• Data For Black Lives - “Data for Black Lives is a movement of activists, organizers, and scientists committed to the mission of using data to create concrete and measurable change in the lives of Black people” [9]
• Decolonial Atlas [10]

Social movements, worker movements

• Algorithmic Justice League [11]
• 'No Tech for Apartheid' and ‘#TechWon’tBuildIt’ campaigns by Tech Workers Coalition
• “The <A+> Alliance for inclusive algorithms is organized by Women@theTable and Instituto Tecnológico de Costa Rica (TEC). It’s a global, multidisciplinary, feminist coalition of academics, activists, technologists, prototyping the future of artificial intelligence and automated decision making to accelerate gender equality with technology and innovation.”

Frameworks for alternative thinking

• Design Justice Network principles
• Data Feminism by D’ignazio & Klein
• Arturo Escobar’s ‘Designs for the Pluriverse’ which argues towards 'design for transitions' and 'autonomous design'
• Pluriversity as a way of teaching and knowing Indigenous methodologies and values.
• Feminist Data Set (Caroline Sinder)- Is each step feminist? Is it intersectional? Does each step have bias? How can that bias be removed?  
• Designing with the future [12]
• Machine consciousness [13]
• Data Justice - eg Couldry & Meijas (2018) or Laura Mann [14]
• “Both/ And” Approaches to knowledge production, to reject notion of binaristic system
• “Walking while asking questions” - Zapatistas motto, democratic potential within decolonial thinking
• Bringing feminist perspectives into community informatics by Peddle, Katrina, Powell, Alison and Shade, Leslie Regan (2008) [15]

Counter-Investigation

Groups using data methods to hold hegemonic powers to account, whilst also reflecting and making aware the biases embedded in such tools.

• Forensic Architecture's methods of constructing political narratives for a place using digital tools to create digital tours and exhibitions, for instance in their project 'Sheikh Jarrah: Ethnic Cleansing in Jerusalem' [16]
• Forensic Architecture's 'Conquer and Divide' interactive web platform to retrace the occupation of the Westbank and Gaza [17]
• Forensic Oceanography [18]
• INTERPRT
• Future predictions and DATA (trend-forecasting) [19]

Education / Pedagogy

• Color Coded LA: ‘We are a POC-only space and collective co-teaching, co-creating, and co-owning new technologies
• Free schooling
• Bermondsey Lampposts case study
• Collective Design School
• Multilingual children’s support [20]
• Criticising Decolonising the University

Critical arts projects:

• Mimi Onuoha’s ‘Library of Missing Data Sets art project
• 'The Work of Art in the Age of Artificial Intelligence: What Artists Can Teach Us About the Ethics of Data Practice' by Kate Crawford [21]
• Media Futures projects [22]
• 'AI Art and Misinformation: Approaches and Strategies for Media Literacy and Fact Checking' by Walker, Thuermer, Vicens, Simperl, 2023

Abolitionist perspectives on technology:

When data / technologies prop up military and policing, it is argued these practices can never be decolonial, so abolitionist principles may be a framework to think through these.

• Carceral Tech Resistance Network: ‘organising against the design, experimentation and deployment of carceral technologies’ 
• The Surveillance Technology Oversight Project: working to end discriminatory surveillance
• Digital freedom, Digital rights, Digital Human Rights discussed by UN Interagency Dialogue on Disinformation and Data Transparency [23]
• Responsible AI [24]
• 'A world without Data' Blog [25]

Co-created resources:

The following pointers are suggested principles and reflective questions for researchers/students to approach decoloniality when working with data methods.

The task of decoloniality is always changing, this is not a definitive set of answers.

Key values that need to be central when approaching challenges:

• Criticality as an approach
• Contextualising and considering the historical background in which data have been selected or extracted and then analysed with personal biases and filters. Thinking in terms of working with data is a reflection about when, how, how data was collected and stored, and who collected it. (Marika/Art practice and research)

Create a loose set of guidelines/principles for arts and humanities projects working with data methods for how to challenge coloniality.

• Elasticity, creating flexible guidelines and principles that can be welcoming and inclusive for under-represented, under-considered data, and their provenance. And also in terms of future thinking, and how to adopt and support the raising of diversity of needs.  
• Developing a set of key values:
  • Reciprocity ← key when thinking about non-extractive data processes
  • Based on social justice principles: access to resources, participation, equity and human rights
  • Equality of privilege - low tech artisanal lived experience data is valued equally with high tech computational big data

What are some critical questions to prompt reflection on projects in relation to decoloniality?

• Why am I doing this?
• For whom am I doing this?
• Who does this serve?
• Who is benefitting from this?
• In what aspects of the project reflect decoloniality?
• Where do they might come from: technical or theoretical problems?
• Whose voice is being heard?

What practical and technical tools do we need?

• (Positionality: Design researcher working with communities)
• Practical tools could range from the decolonial ideologies of working ‘with’ and not ‘for’. Unlearning: decolonising means a lot of unlearning and learning, where there might be a completely flipped version of accessing and transferring data.
• Intersectionality.
• Empathy
• Certain level of cultural, socio-political understanding and linguistic ability of the topic of research

What perspectives, thinking tools, frameworks, theories do we need to engage in decoloniality?

• Pluriversal / pluriverse / pluriversal thinking
• Both / And
• Perspectives - accessible intros to standpoint epistemology, intersectionality, the gaze, etc that can tap into people’s lived experiences
• Real engagement with data itself and data tools so that conversations can move from critical ‘hovering over’ of the academic eye to questions around empowered ways to work with data – not just as a data subject, but a ‘data agent’ (not sure what it should be called because in reality we become ‘data workers’ which is not necessarily at all empowering. So there seems like some tension between acquiring data skills as empowering and becoming a data worker as falling back into less agential roles. Or something like this!
• A side note on LLMs (large language models) and ethics
• Conversational generative AI such as ChatGPT, Bart etc offers a really exciting opportunity for the co-analysis of qualitative data. However, ANYTHING you put into a GPT interface becomes training data for that product. It is therefore crucially important to get INFORMED CONSENT from anyone whose contributed data (anonymous or not) you intend to put into a chatbot. This informed consent includes telling them that their data will be retained and used as training data.They may be fine with this! But they need to know.
• You may also find that your institution’s data protection rules require you to declare if you are sending data outside of countries with EU GDPR parity. Open.AI’s servers are based in the US.
• Data Justice Plan Template This was created by Johanna Walker & Gefion Thuermer after we had worked on several EU projects which used cascading funding to support small to medium enterprises to develop smart cities products and services using citizen data. Where data is silently and invisibly collected by sensors and manipulated by algorithms, civil society is at risk of being simultaneously used as the subjects and objects of data, yet with no guarantee of receiving its benefits. This is a form of data injustice, where the benefits of data flow in one direction and are not evenly or fairly distributed. Data that is collected with citizens as its subject or object  should benefit the community that it came from. Think about how you can ensure that this happens for your project data.
  • Who will contribute data to your project/artwork, and why?
  • How will they benefit from what you plan to do with the data?
  • How will this benefit materialise?
  • Who will be responsible for making this happen?
  • What will happen if the project/artwork does not go to plan – what can participating communities benefit from partial results, or what additional results might become available?
  • Who else will benefit from the data and how? We have also been testing this template with data artists.