Algorithmic Settler Colonialism: The Impacts of Artificial Intelligence on Indigenous Data Sovereignty in Australia
I INTRODUCTION
Colonialism in Australia has never been confined to the occupation and destruction of land. Since the nineteenth century, the imposition of Western authority over Indigenous territories, laws, and modes of production has been accompanied by colonial control over all aspects of Indigenous cultures, languages, and knowledge. Imperialistic power structures are detrimental to the preservation of First Nations sovereignty, as Indigenous peoples possess a complex, relational connection to kinship and Country across their spiritual, sentient, and physical worlds which are deeply rooted in ontologically and epistemologically divergent networks. These networks are vastly different to Western conceptions of knowing and being, as they align and shape cultural identities through ‘Indigenous laws, languages, and protocols determined by the nature of Country itself’. Colonialism in Australia consequently extends beyond physical dispossession, as there are ongoing and implicit epistemological issues regarding what qualifies as legitimate knowledge and how this knowledge is used. The positional authority that colonial norms occupy over Indigenous peoples’ lives is systemic and has now expanded into the digital realm. As Artificial Intelligence (‘AI’) technologies become increasingly prevalent, concerns naturally arise regarding the norms it encodes, whose interests its outputs serve, and which institutional hierarchies are entrenched.
In this essay, AI is understood as a broad category of data-driven technologies, though the primary focus will be on predictive and decision-making systems, including algorithmic decision-making and predictive analysis tools. Since these systems operate with less visibility and are progressively embedded within public systems such as policing and education, it is important to understand the implicit biases that exist in their functions. AI cannot be framed as a neutral tool so long as its development remains largely constrained by Western epistemologies that determine how information and knowledge are defined, measured, and operationalised. This issue is encapsulated in the concept of algorithmic settler colonialism: that is, the manners in which digital technologies – particularly AI and algorithmic systems – ‘reinforce and extend settler colonial structures of power, control, and dispossession’. This essay contends that existing AI technologies perpetuate algorithmic settler colonialism through digital systems that reinforce colonial power structures and discriminatory biases, thereby erasing Indigenous peoples’ right to self-determination and sovereignty. Part II establishes the historical and theoretical context of algorithmic settler colonialism. Parts III and IV discuss the impacts of AI on racialised surveillance and the erasure of Indigenous cultural practices. Finally, part V situates Indigenous Data Sovereignty as an imperative ethical framework grounded in ‘relational accountability, cultural integrity, and collective rights’ to enable self-determined AI governance.
II HISTORICAL AND THEORETICAL FRAMEWORKS OF ALGORITHMIC SETTLER COLONIALISM
Since its inception in the 1950s, the field of AI has been inextricably linked with colonial logics of domination, which reflects and reinforces enduring corrupt power structures. The 1956 Dartmouth Workshop, which marked the initial conceptualisation of AI, took place within a sociopolitical context heavily influenced by anxieties around the Cold War and escalating military-industrial complex. During this conference, a strong emphasis was placed on optimising problem-solving as the core purpose of digital intelligence, which has since ‘directed AI’s trajectory toward applications that often prioritise control, efficiency, and strategic advantage’ to align with the interests of those in power. Early designs of AI such as statistical inference and symbolic manipulation emulated dominant (Western) computational models and were swiftly implemented for military applications, intelligence gathering, and industrial automation. This alignment with and service to established power structures from the outset renders existing AI systems incapable of processing knowledge that is collective, relational, and place-based in a manner consistent with Indigenous culture.
In the digital age, settler colonialism manifests through data extraction and AI-driven decision-making, resulting in the erasure of Indigenous knowledge, laws, languages, and culture. Since AI systems are often trained on datasets carrying colonial biases, Indigenous perspectives are consequently misrepresented or omitted altogether due to distorted narratives, epistemic violence, and continued digital exclusion. This essay embraces critical and decolonial legal theory, framing coloniality as a ‘structure of power that survives the formal end of colonial rule’ and recognises its continuance in the organisation of law, knowledge, and institutions in postcolonial states. Western knowledge technologies – specifically ‘sorting systems and those created from statistical sciences’, such as AI algorithms – are inextricably linked to colonial logics of order, as they are designed using structures and ideologies that impose an imperialistic positional authority. This positionality reinforces the West’s view of itself as the centre, arbiter, and source of legitimate and “civilised” knowledge.
Most of the existing data that are accessible via the internet and related algorithms are based on predominantly White, neoliberal, heteronormative, settler colonial assumptions of what is “natural”. Subsequently, systemic biases such as the conflation of race and poverty with crime and recidivism are encoded into digital systems built around a Eurocentric ethos of optimisation, efficiency, and profitability. Researchers have found that Indigenous peoples are particularly prone to in-built digital biases, as generative AI systems describes them in a stereotypical and reductive manner approximately 80% of the time, whilst minority groups generally are ‘hundreds to thousands of times more likely to [have] their identities [portrayed] in a subordinated manner’. Subsequently, AI does not only reflect the unjust dynamics of society; it actively reinforces the outdated assumptions and values of those in power and has the potential to exacerbate inequality without enforceable governance structures directly informed by Indigenous perspectives.
III RACIALISED SURVEILLANCE IN AUSTRALIA’S CRIMINAL JUSTICE SYSTEM
One of the most dangerous implications of algorithmic settler colonialism is its reproduction of racial hierarchies, whereby AI systems ‘operationalise and extend forms of structural racism that are historically grounded in colonial regimes of classification, surveillance, and control’. Oftentimes, these systems are trained on datasets drawn from colonially structured archives or biased digital traces, consequently embedding racism, sexism, and ableism into the foundations of automated systems. Such encoded biases are a form of digital colonialism, where platform infrastructures and data extraction entrench colonial logics, reinforcing asymmetries of power between those who generate data and those who control and profit from it. Given the overrepresentation of Indigenous peoples in Australia’s criminal justice system, algorithmic decision-making only exacerbates existing forms of discrimination through racial profiling and over-policing. One example can be found in facial recognition technologies, which ‘consistently produce higher error rates for Black and Indigenous peoples’, including misidentification of suspects or perpetrators, which creates intensified forms of racialised surveillance and policing. These predictive tools have been criticised for perpetuating settler colonial control while failing to resolve the structural inequalities that contribute to criminalisation.
New South Wales Police’s ‘Suspect Target Management Program’ further illustrates the notion of racialised surveillance stemming from algorithmic settler colonialism. Before its discontinuation in 2023, this algorithm-based tool identified a disproportionately high number of Indigenous peoples as being at high risk of committing criminal offences. The Australian Human Rights Commission found that a common problem with these AI systems is that certain racial or ethnic groups, particularly Indigenous peoples, may experience excessive scrutiny by police over a long period of time, which necessarily ‘contributes to members of that group being convicted at a disproportionate rate for particular crimes’. When such biased statistics are used to train datasets for risk assessment tools, these tools will incorrectly ‘learn’ that individuals from particular marginalised groups are more likely to commit crimes rather than recognising the data as resulting from past discrimination.
IV ERASURE OF INDIGENOUS ART AND CULTURE
Beyond the criminal justice system, algorithmic settler colonialism threatens Indigenous sovereignty over linguistic and cultural practices. For instance, in Aotearoa (New Zealand), Māori have raised concerns that AI is reproducing versions of their language without Māori governance in ways that could result in Indigenous communities losing sovereignty over their own language. In Australia, there are similar concerns regarding the Indigenous art industry, particularly AI-generated “art” that mimics First Nations motifs and styles without cultural permission and attribution. The appropriation and commercialisation of Indigenous art has been a longstanding issue and, evidently, AI has only amplified them. Since these systems are trained on vast datasets often sourced from the internet without consent, there is an increasing risk of Indigenous artists losing control over their cultural expressions, designs, and symbols. This is not only a matter of copyright infringement but of cultural sovereignty and self-determination.
The commodification of Indigenous art in digital spaces poses a significant threat to cultural knowledge preservation, as it positions Indigenous art as ‘reproducible by and accessible to anyone’. This issue reflects what Aileen Moreton-Robinson refers to as a logic of ‘white possession’, where sacred cultural practices are converted into ‘a form of property to be commodified and privatised’, further hindering Indigenous peoples’ authority to control ‘datafied representations of [their] knowledges and peoples’. Given that Indigenous lore and law have historically been communicated through ‘poetic traditions of rock-wall, body and sand painting, carving, song, and performance’, AI-generated “art” is an appropriation of Indigenous identity and cultural expression, particularly where such identity has been forged with resilience in the face of racism and adversity. For many Indigenous artists, their creative practices are also a form of cultural activism: within Australia’s postcolonial context, the creation and distribution of Indigenous art can itself constitute a political assertion of sovereignty and resistance. AI systems that reproduce First Nations motifs without permission therefore not only violate individual artistic ownership, but also the cultural and political agency that is inextricably tied within Indigenous creative expression. As AI technologies become increasingly integrated into content generation, it is imperative that Indigenous peoples have the power to assert their digital sovereignty to prevent further discrimination, misrepresentation, misappropriation, and loss of control over their cultural legacies.
V INDIGENOUS DATA SOVEREIGNTY
While this essay interrogates the systemic biases reproduced through algorithmic settler colonialism, it is equally important to acknowledge the benefits AI technologies have delivered for First Nations communities. These benefits provide an informative foundation for Indigenous Data Sovereignty (‘IDS’) principles and practices that AI companies should adopt in their ethos. In Australia, environmental rangers are using generative AI to ‘car[e] for Country’ by converting oral reports into written formats as required by government guidelines. Furthermore, traditional owners in Kakadu have been using AI to streamline and maximise their capacity for conservation management across millions of hectares. Learning accessibility in higher education has also improved as students are able to use AI technologies to practice First Nations languages and refine their notes, therefore reducing linguistic barriers. While AI has its advantages, researchers caution that current frameworks risk eroding data sovereignty; therefore, Western AI corporate ethics and reconciliatory mechanisms will not suffice. There must be total structural realignment through Indigenous-led frameworks such as IDS.
Walter describes IDS as concerning ‘the rights of Indigenous peoples to own, control, access and possess data that derive from them, and which pertain to their members, knowledge systems, customs or territories’. Grounded in the ‘inherent rights of self-determination’ as stipulated in the United Nations Declaration on the Rights of Indigenous Peoples, IDS is a growing movement that advocates for Indigenous governance and oversight for the ‘entire knowledge digitation process’. Miam nayri Wingara (an IDS collective in Australia), in partnership with the Australian Institute of Indigenous Governance, highlighted the importance of IDS in ensuring First Nations communities’ control of data ecosystems, including their development, stewardship, dissemination, and infrastructure. The collective also advocates for adherence to the CARE principles – Collective Benefit, Authority to Control, Responsibility, and Ethics – which is a community-centred framework that ensures IDS goes beyond Western proprietary notions of access and ownership. Subsequently, IDS represents a contumacious epistemology compared to Western conceptions of digital rights, as it not only exceeds ‘narrow notions of sovereignty as data localisation and jurisdictional rights, but also upends the assumption that the nation state is the legitimate locus of power’.
Departing from ‘a white-constructed form of Indigenous proprietary rights’, IDS centres Indigenous self-determination over the entire lifecycle of Indigenous data, which requires the abolition of positivist relations to data and rejection of assumptions that they are objective or neutral and thus not in need of culturally relevant stewardship. For example, researchers found that rule-based systems are most compatible with IDS principles, as their transparent, logic-driven design enables direct embedding of Indigenous cultural conventions, legal orders, and ethics. Accordingly, rule sets can be authored and reviewed by Indigenous authorities, aligning with ‘community norms and governance’ across the whole data life cycle. By structurally embedding Indigenous ‘agency, worldviews, and data priorities across diverse data systems’, there will be greater transparency and accountability for AI’s development and deployment.
IDS practices are slowly emerging across Australian jurisdictions. In Victoria, the Yoorrook Justice Commission employs data sovereignty principles and ‘advocates against state copyright over Indigenous data produced in Royal Commissions’. Yoorrook’s 2022 Interim Report outlines its efforts to ‘investigate, document, and recommend…reform for systemic injustices experienced by First Peoples in Victoria since the start of colonisation’. While their IDS framework has not been expressly adopted in Australian AI systems, Yoorrook’s model offers a blueprint for future Indigenous data governance frameworks: it is designed and led by Indigenous Elders and peoples, ‘conducted in line with First Peoples’ ways of knowing, being and doing…and empowered to investigate any systemic injustice it considers appropriate’. It is also the first State inquiry to apply IDS principles to its work, which, in practice, is a step towards systemic change outside of data governance and digital networks.
Currently, Yoorrook’s realisation of IDS is constrained by federal law which requires ‘all Royal Commission records…to be transferred to the Public Records Office of Victoria when the Commission’s work has concluded’. This requirement is being actively protested as expressed in Yoorrook’s interim report: ‘the government [must] urgently progress the necessary legislative changes to enable the implementation of First Peoples’ choices about how the information they provide to Yoorrook is to be stored, accessed and used in the future’. Evidently, the actualisation of IDS can have broader implications beyond its initial scope as a project or organisation’s initiative by generating momentum in the data justice movement, increasing pressure on governments and corporations to centre Indigenous knowledge, interests, and sovereignty. This movement facilitates a reframing of AI technologies from Western ideals of efficiency and optimisation to recognising that AI is part of an ‘interconnected system with ethical obligations to the communities and lands it affects’, with communal benefits being at the core ethos. Overall, IDS provides a culturally-informed, First Nations-led governance framework for AI systems where its development and deployment prioritises ‘relationality, responsibility, and reciprocity’.
VI CONCLUSION
Carlson and Worrell perfectly encapsulate the crux of this essay: ‘the question is not whether machines can be controlled, but whose interests they are programmed to serve, and at whose expense’. Existing AI technologies reflect a world structured by inequalities, and these systems operate according to the sociopolitical and economic logics they are trained within. Such outdated logics are grounded in settler colonialism, racial capitalism, and extractive economies. From algorithmic bias in the criminal justice system to the exploitation and appropriation of Indigenous art, AI continues to ‘scale historical injustices under the guise of neutrality and progress’. Indigenous scholars have highlighted the urgency for First Nations communities to shape the development of AI in ways that centre ‘relationality, sovereignty, and Indigenous aspirations’, ensuring activate participation and protection of cultural integrity and digital justice in the AI landscape. Therefore, it is imperative to interrogate and defy the underlying imperialistic power structures, assumptions, and values that are digitally encoded, especially as AI becomes increasingly entwined with the governance of everyday life.
BIBLIOGRAPHY
Articles/Books/Reports
Abdilla, Angie et al, ‘Out of the Black Box: Indigenous Protocols for AI’ (Research Paper, Deakin University, 2021) 2
Australian Human Rights Commission, Thematic Report on Artificial Intelligence (AI) and Racial Discrimination (Submission to Special Rapporteur on Contemporary Forms of Racism, March 2024)
Coleman, Beth, ‘Human-Machine Communication, Artificial Intelligence, and Issues of Data Colonialism’ in Andrea Guzman, Rhonda McEwen, and Steven Jones (eds), The SAGE Handbook of Human-Machine Communication (Sage, 2023) 350
Du, Shuili and Chunyan Xie, ‘Paradoxes of artificial intelligence in consumer markets: Ethical challenges and opportunities’ (2020) 129 (1) Journal of Business Research 961
Ertel, Wolfgang, Introduction to Artificial Intelligence (Springer, 3rd ed, 2025)
Fisher, Laura, Aboriginal Art and Australian Society (Anthem Press, 2016)
Kukutai, Tahu, ‘Indigenous Data Sovereignty – A New Take on an Old Theme’ (2023) 382(6674) Science 1
McDaniel, John and Ken Pease, Predictive Policing and Artificial Intelligence (Routledge, 2021)
McMaster, Natalie, ‘Generative AI and Australian First Nations Representation: Ethical Concerns and Cultural Implications’ 2026 11(1) Australasian Journal of Technology Education 1
Mehrabi, Ninareh et al, ‘A Survey on Bias and Fairness in Machine Learning’ (2021) 54(6) ACM Computing Surveys 1
Miller, Bronwyn, ‘Colonial Legacies and the Settler Colonial Present in Google: An Anticolonial and queer Response’ (PhD Thesis, University of New South Wales, 2026)
Moreno, Jonathan et al, ‘The ethics of AI-assisted warfighter enhancement research and experimentation: Historical perspectives and ethical challenges’ (2022) Frontiers in Big Data 5
Moreton-Robinson, Aileen, Talkin’ up to the White Woman: Indigenous Women and Feminism (University of Minnesota Press, 2021)
Morris, Alex, ‘The Dilemma of Digital Colonialism Unmasking Facial Recognition Technology and Data Sovereignty in Aotearoa New Zealand’ (2023) 19(1) Police Quarterly 80
Phillips, Patrick and Nicholas Ng-A-Fook, ‘An Unsettling Artificial Intelligence: Algorithms, Curriculum, and Futurities’ (2024) 16(1) Journal of the American Association for the Advancement of Curriculum Studies 11
Prehn, Jacob et al, ‘Indigenous Data Sovereignty and the Governance of Artificial Intelligence: Toward Equitable Benefit for Indigenous Peoples’ (2026) (Winter) Journal of Sociology 2
Sloane, Mona, ‘Controversies, contradiction, and ‘participation’ in AI’ (2024) 11(1) Big Data & Society 1
Smith, Linda, Decolonizing Methodologies: Research and Indigenous Peoples (Bloomsbury Academic & Professional, 3rd ed, 2021) 71
Walter, Maggie et al, Indigenous Data Sovereignty and Policy (Routledge, 1st ed, 2020)
Worrell, Tamika and Bronwyn Carlson, ‘Indigenous AI Futures: Uncle Chatty Gee, Aunty Lexi, and Algorithmic Settler Colonialism’ (2025) 15(3) Somatechnics 296, 303
Worrell, Tamika and Bronwyn Carlson, ‘Robots Behaving Badly: Algorithmic Colonialism and the Consequences of AI’ (2026) (Winter) Journal of Sociology 1
