As African governments develop national artificial intelligence strategies, questions of safety, economic competition and employment have dominated the policy debate.
Less attention has been paid to how AI expansion may affect environments, water and biodiversity, or the rights of communities whose territories and knowledge are being encroached upon.
These concerns arise at several points in the AI value chain. The associated data centres require energy, water and land, while their hardware depends on mineral supply chains.
AI systems may collect and reproduce cultural, linguistic and environmental knowledge without recognising that such information is often held collectively by a community, rather than by an individual. This is important because conventional data protection frameworks tend to focus on individual data privacy and ownership, making them poorly equipped to protect collectively held information.
In July, the UN held its first-ever Global Dialogue on AI Governance, in Geneva, Switzerland. Among those in attendance was Hindou Oumarou Ibrahim. She is president of the Association for Indigenous Women and Peoples of Chad, and a former chair of the UN Permanent Forum on Indigenous Issues. While there, she took part in discussions on advancing environmentally sustainable AI governance that also respects the rights of Indigenous peoples.
Here, Ibrahim tells Dialogue Earth about the importance of bringing Indigenous perspectives to Africa’s AI governance debate. She explains why African governments should incorporate biodiversity, land rights and collectively owned data considerations into their AI policies. She says Indigenous peoples must be recognised as rights holders and decision-makers – especially when newfangled technology has implications for their lands or intellectual heritage.
The interview has been edited for length and clarity.
Dialogue Earth: As African countries develop national AI strategies, what opportunities do governments have to reflect the continent’s environmental realities and cultural diversity?
Hindou Oumarou Ibrahim: Africa is home to extraordinary biological and cultural diversity, but it also faces significant pressures from climate change, mineral extraction, land degradation and infrastructure expansion. National AI strategies should therefore connect technological development with biodiversity protection, climate resilience, land rights and the responsible management of natural resources.
African governments can lead by requiring environmental and social impact assessments for AI infrastructure. This would help protect water and land from unsustainable data centre development, and ensure that mineral demand does not drive ecosystem destruction or dispossession.
They can also make Indigenous peoples formal participants in AI governance and protect collective rights over our rich and diverse cultural, linguistic and environmental data.
There is also an opportunity to invest in African and community-controlled AI capacity: local research institutions, Indigenous peoples’ innovation centres, African language and culture models, and technologies designed for African and community needs, rather than external markets.
What are current AI governance debates missing by leaving biodiversity and Indigenous rights at the margins?
AI governance has largely been built around a western, human-centred understanding of risk. Policymakers are rightly concerned about issues such as misinformation, employment, discrimination and security. But they have paid far less attention to the natural systems that make economies and societies possible. As a result, biodiversity loss, ecosystem damage and Indigenous peoples’ rights are rarely treated as core AI governance issues.
There is also a wider risk. Without ecological guardrails, AI could accelerate extraction, consumption and infrastructure development in ways that place even greater pressure on ecosystems and Indigenous peoples. Many of the lands under pressure from the mineral, energy and infrastructure sectors are Indigenous peoples’ territories and biodiversity-rich areas.
AI governance cannot be described as responsible if it does not protect the ecosystems and cultures on which communities depend.
What should meaningful Indigenous participation in AI governance involve?
Indigenous peoples should help shape the safeguards that protect their rights, knowledge and territories. They should participate in decisions about how systems affecting them are deployed. This must be done with respect for free, prior and informed consent.
FPIC is about allowing people to give or withhold consent for projects that may affect them. Its main and original focus is Indigenous peoples but it is also considered best practice for local communities of all kinds. FPIC is enshrined in the UN Declaration on the Rights of Indigenous Peoples, the Convention on Biological Diversity, and the International Labour Organization Convention 169.
Participation is not only about consultation. Indigenous peoples must be recognised as decision-makers and rights holders, particularly when AI policies, infrastructure projects or data practices affect their land, knowledge or cultural heritage.
How are existing data protection systems failing to protect knowledge and data held collectively by communities, rather than individuals?
Conventional data protection systems tend to focus on individual privacy. They do not adequately protect a community’s collective rights over its language, cultural heritage, traditional knowledge or territorial information.
Indigenous peoples must retain control over whether and how their knowledge and data are used. Governments and companies should recognise that Indigenous peoples’ data is often collective and develop consultation frameworks that reflect this.
AI companies should also audit their models for linguistic and cultural bias, include Indigenous leaders and experts throughout the process, and create mechanisms that allow communities to challenge or remove inaccurate or harmful representations.
Can AI support Indigenous knowledge and conservation without replacing community expertise?
There are encouraging examples in which AI is supporting, rather than replacing, Indigenous peoples’ knowledge. Some examples are outlined in a recent report to the UN Economic and Social Council on Indigenous peoples and AI.
In the Amazon, Indigenous communities have worked with NGOs and researchers to use AI-supported satellite monitoring to detect illegal deforestation and mining. In Chad, participatory mapping and predictive analysis have been combined with pastoral knowledge to anticipate drought and protect herder’s migration routes. In the Arctic, Sámi and Inuit communities have integrated traditional knowledge with predictive models to monitor changes in sea ice, wildlife movements and the impacts of extractive industries.
The important point is that the technology should strengthen existing knowledge rather than replace it. Indigenous peoples bring detailed, place-based understanding accumulated over generations. AI can help analyse information and identify patterns, but it cannot substitute for the cultural context, relationships and judgement that make that knowledge meaningful.
AI can also help communities and conservation organisations process large volumes of information more quickly. It can support the detection of deforestation and illegal extraction, identify changes in ecosystems, analyse satellite or acoustic data, predict environmental threats and improve monitoring in places that have historically lacked resources.
The strongest opportunities are those in which AI supports priorities defined by communities themselves and strengthens the capacity of people already protecting nature.
But these benefits are not automatic. Technology can reinforce inequality when tools are designed without community participation, when data is extracted without consent, or when communities lack the infrastructure and funding needed to use the systems themselves.
Using AI to monitor forest loss does not compensate for an AI growth model that drives greater energy use, mineral extraction, water consumption or habitat destruction.
What is one AI governance recommendation you would make to world leaders?
Leaders must ensure that no AI system, policy or infrastructure project affecting Indigenous peoples, their data or their territories proceeds without their free, prior and informed consent. That principle should apply across the entire AI value chain: from the extraction of minerals and construction of data centres to the collection of Indigenous peoples’ knowledge, and the deployment of AI tools in communities.
This is not only about preventing harm. It is about recognising Indigenous peoples as decision-makers and rights holders. World leaders should embed Indigenous peoples’ participation and data sovereignty directly into national and international AI governance frameworks, rather than treating them as optional social safeguards.
