Every five years, Australia conducts the Census – the national snapshot revealing who we are, how we live and what we need.This year, it will be held on August 11.Governments rely on it to plan hospitals, schools, transport and housing, while researchers use it to examine how the country changes. Read more: 85% of Australians will complete the census online. Here’s how to avoid potential scams For Aboriginal and Torres Strait Islander peoples (First Nations), it influences decisions that directly affect funding and essential services.First Nations and the CensusThe Census has a complicated history with First Nations peoples.Until 1967, First Nations peoples were excluded from being counted.Today, many First Nations communities question how their information is collected, used, who controls it, and if it ultimately supports or harms them.Recognising Australia is evolving, new questions will be in the 2026 Census. It will include questions on sex recorded at birth, gender identity and sexual orientation and expanded ancestry options. Read more: For the first time, the census will ask your sexual orientation. My new study shows why it matters However, the Australian Bureau of Statistics said it needed longer to consult with First Nations community representatives before including a potential new question about First Nations’ cultural identity in the Census.The decision to not include the potential new question will further erode many First Nations peoples’ trust in national data collection.Why the distrust?A recent Murray-Darling Water and Environmental Research Program project completed by the One Basin Cooperative Research Centre explains ethical issues with use of public datasets like the Census.It also outlines a path for improvement, showing how to build ethical, accurate, community-led First Nations data systems that provide Indigenous data governance.Although Census data appear neutral, they are shaped by the values and assumptions of those designing the questions and interpreting the results. Without First Nations involvement in the design, standardised forms such as the Census fail to reflect the cultural, linguistic, or household realities of First Nations peoples. Our research found public datasets often treat First Nations information as free to access, use and interpret without community consent. This renders it “ownerless” and available for use by governments and researchers without First Nations input.We refer to this as “data nullius”, echoing the colonial fiction of “terra nullius”, which wrongly claimed land was ownerless before colonisation.Long histories of data misuse and government intervention means many communities approach the Census with caution. Beyond the Census, records such as birth and death registers, land titles and research files often left out or misrepresented First Nations peoples, turning complex cultures into negative labels.These archives were used by governments to control and monitor First Nations peoples, as well as to record their dispossession.Remote communities also face logistical challenges such as distance, weather and short collection windows, while mobility between households, homelessness, and overcrowding further complicate Census counting.The move to mainly online responses – respondents can still request a paper form – can also hinder communities with limited or unreliable internet access. When these structural issues combine, First Nations people remain misrepresented. What we researchedAustralia, like Aotearoa (New Zealand), Canada and the United States, lacks strong protections for First Nations data. We analysed 30 Australian government agencies, universities and research organisations, investigating whether they had policies regarding First Nations data. We searched for policies on Indigenous data sovereignty, rights, consent, benefit-sharing, cultural authority and transparency.Findings were stark:less than half had enforceable protections in placestate agencies had the weakest safeguardsmany organisations used language of partnership but lacked mechanisms for consent or accountability.In short, many Australian data systems holding First Nations information are not yet built for or by First Nations.As part of our research, we extracted and consolidated a set of 27 Indigenous data governance principles. These were drawn from published literature, alongside national and international best-practice exemplars. These principles are designed to identify where ethical standards diverge from current public data practices. They must be centred around Indigenous data sovereignty and grounded in country, culture, community, relationality and the need to care for future generations.Rather than imposing a fixed checklist, these principles establish an agile framework. They can be assembled into flexible diagnostic tools, adaptable to context and evolving practice. Persistent undercountingA major issue in this space is the ongoing undercount of First Nations people.In 2021, the Australian Bureau of Statistics estimated a 17.4% undercount of around 170,800 First Nations people missing from the national dataset. The figure was almost identical in 2016.First Nations peoples are among the most analysed populations in the world, but rarely the ones who control the information. This creates what scholars call the “Indigenous data paradox”: large volumes of data about First Nations peoples but very little data for or by them. When communities are unseen or misrepresented, trust erodes and Census participation declines, along with it the government’s ability to deliver essential resources.Data as a relationship, not a resourceFor many First Nations peoples, data are not merely numbers. They are story, identity, kinship and connection to Country. Handled without care, data harm. But when First Nations data are governed by communities themselves, they become tools for visibility, self-determination and better decision-making. In the right hands, data have power: they can secure rights, direct resources, protect culture and shape futures.Australia needs to transform how it collects and uses First Nations Census data. It should do this by co-designing questions, embedding Indigenous governance and sovereignty, addressing digital divides, and using culturally grounded counting. Results must also be returned in ways communities recognise, trust and benefit from.Birri O'Dea works for the One Basin Cooperative Research Centre – its researchers receive funding from the Australian government.Chesley Cutler works for Griffith University.Troy Meston works for the One Basin Cooperative Research Centre – its researchers receive funding from the Australian government.