For Indigenous communities, AI brings peril — and promise

When the United Nations marked the International Day of the World’s Indigenous Peoples last week, it signaled a growing recognition of a new kind of extraction. Artificial intelligence, or AI, systems are being trained on massive troves of online data, much of it collected without the consent of the communities involved. For Indigenous peoples, this new form of extraction has raised questions about who controls their histories, languages, and cultural knowledge and whether the technology will erase or distort them entirely. With this in mind, tribes and nations have been pushing to assert “data sovereignty” — the right to control how information is collected and used — and claim a seat at the table as tech companies and governments set the rules for AI oversight. 

Perhaps as a reflection of the AI boom and its implications for Indigenous communities, this year’s theme for the international Indigenous day was “Indigenous Peoples and AI: Defending Rights, Shaping Futures.” At the heart of the commemorative event was the acknowledgement of a stark reality: While information extraction, data center development, and the global critical minerals race have harmed Indigenous communities, emerging AI technologies also offer tools to address some of these challenges. The critical minerals used in batteries for backup power at data centers are often sourced from ecologically sensitive regions in Indigenous territories such as the Atacama Salt Flat in Chile. The centers also consume vast amounts of water and energy, which can strain local aquifers and cause pollution. But even as communities grapple with ecological destruction and changes to their way of life due to the AI boom, some are beginning to recognize the potential it holds for language revitalization and climate forecasting.  

“Indigenous people have a long history asking to be seated at the table and to ensure that those who have the power listen to their demands,” said Fernando Marani, program director of justice, inclusion, and equality at New York University, during the event. Those who currently hold power “are not governments, but the private sector,” he added. 

VEJA  FloridaCommerce now accepting applications for state funding to enhance military communities

Panelists from Norway, Canada, New York City, Thailand, and the Amazon discussed how AI affects the rights and livelihoods of Indigenous communities during the event. By adopting a rights-based framework, such as the principle of free, prior, and informed consent, tribes and nations can assert data sovereignty and drive the direction of AI development and consultation efforts, they said. 

Tech companies train their generative AI and models such as ChatGPT, Claude, X’s Grok, and Google’s Gemini on vast quantities of text, images, videos, and other online data. The information is typically used without shared agreement or approval. Some of the datasets used to train these models come from governments that historically undercount, miscount, or otherwise fail to accurately represent Indigenous communities. Indigenous people, for instance, are routinely misclassified on death certificates. More recently, a Canadian television network identified a number of Indigenous language learning books with inaccurate translations likely produced with AI tools. These processes are “threatening to distort and erase our knowledge, rather than preserve it,” said Danielle Boyer, an Anishinaabe robotics inventor and founder of Skobots. 

Read Next


Data centers, drought, and dispossession: The real nightmares in Ari Aster’s ‘Eddington’

At the same time, Indigenous communities are beginning to harness AI to address their own priorities, from combating climate change to revitalizing endangered languages. PolArctic, cofounded by Yup’ik Alaska Natives Leslie Canavera and Lauren Decker, develops AI forecasting tools in conjunction with Indigenous knowledge. These tools largely help fishing businesses plan sea ice forecasts and predict fish stocks in the Arctic Ocean. Using shared knowledge of the best fishing spots, the company practices a balance between open data principles and Indigenous data sovereignty.

Similarly, Te Heku Media, a Māori broadcasting organization in New Zealand, built a speech recognition tool to transcribe te reo, the Māori language. The broadcaster trained on data using hours of native speaker audio and consulted with the Māori community. Eventually, the organization built a distribution platform known as Whare Kōrero, or “house of speech,” to store the data used to train the language model.

VEJA  On Controlling Fire, New Lessons from a Deep Indigenous Past

“The data generated from tribal communities, or even from tribal citizens being internet citizens, and signing up on social media apps, or just participating in the internet more broadly, all are a commodity,” Rusty Pickens, a member of the Chickasaw Nation, told Grist. Pickens previously worked in the State Department and White House on new media technologies and digital platforms. “How do we protect our communities and make sure you know we’re not exploited and our resources aren’t extracted yet again because of novel technology that’s coming out?”

There are a few regulations governing AI. In 2024, the United Nations adopted the Global Digital Compact, the first international framework guiding AI governance. However, the compact currently doesn’t include protections for Indigenous data. The European Union AI Act, which passed last year, also aims to minimize discrimination and bias in AI use and “set a strong step forward,” according to Anne Torill Nordsletta, a director at the Norwegian Centre for E-Health Research. “But true inclusion means more than safeguards. It means making sure Indigenous peoples are not invisible in the data used to build AI,” she said.

Pickens pointed out that AI’s rise bears similarities to coal, oil, and gas, and more recently mineral extraction on tribal lands. 

“We were moved off of our tribal homelands east of the Mississippi River for land purposes, and we got to Oklahoma, then they discovered oil. Now we’ve had a land run on all the mineral rights,” he said. “Those extractive cycles of history seem to keep repeating themselves. AI and our data are another iteration of that, and I’m hoping we’ve learned our lessons from the past, and tribes can assert themselves and come to the table.”


Postagem recentes

DEIXE UMA RESPOSTA

Por favor digite seu comentário!
Por favor, digite seu nome aqui

Stay Connected

0FãsCurtir
0SeguidoresSeguir
0InscritosInscrever
Publicidade

Vejá também

EcoNewsOnline
Privacy Overview

This website uses cookies so that we can provide you with the best user experience possible. Cookie information is stored in your browser and performs functions such as recognising you when you return to our website and helping our team to understand which sections of the website you find most interesting and useful.