News Release

AI could be the future for preserving marginalized cultures, say experts

Book Announcement

Taylor & Francis Group

Promising new AI tools are being developed to protect and preserve the cultural traditions and heritage of marginalized or indigenous communities, including language, folklore, oral traditions, and community wisdom.

These advances have prompted industry experts to highlight the potential for culturally aware AI systems, which can ensure unique traditions are not only remembered but also shared and reimagined across generations.

But to continue this positive trend, tech companies must ensure the AI-based systems they develop are representative and build bridges across cultures, experts say.

In the forthcoming book AI for Community, a group of technologists, researchers, and cultural advocates explore how AI can preserve heritages that might otherwise be at risk of being lost. The authors highlight how AI, when built in collaboration with communities, can help carry forward traditions across generations.

“This technology is opening up new ways to capture and share cultural knowledge,” says co-author and AI strategist Iran Davar Ardalan. “We’re seeing AI used to record endangered languages, create interactive learning tools, and even preserve oral histories, offering people new ways to connect with their roots. That kind of connection can inspire pride, belonging, and new creative expression.”

She adds, “From Austin to Aberdeen and Athens to Adelaide, every region holds its own wisdom. These are passed down through lullabies, idioms, recipes, and stories. When AI is informed by these traditions, it does more than process information. It helps carry forward what communities already know and live by. By grounding AI in local context, we can create smarter, more adaptable products that truly meet the needs of the people who use them.”

Protecting language

The authors encourage AI developers to think beyond functionality and consider how the systems they build can reflect different cultural roots, traditions, and values. Rather than applying a one-size-fits-all approach, the multicultural team of authors urge technologists to work closely with cultural experts and communities to develop tools that can adapt to the regional traditions and perspectives of each community.

One example is the collaboration between Howard University and Google, and the creation of speech datasets representing African American Vernacular. Though these datasets haven’t yet been integrated into products, these resources offer huge potential for improving voice recognition systems for African American people, and ensuring the systems respond accurately.

AI has also been developed to protect native languages at risk of extinction. UNESCO estimates that 3,000 languages are in danger of completely disappearing within the next century – the equivalent of one every two weeks.

A language that is at-risk is Māori, but Te Hiku Media in New Zealand has developed a pioneering AI model aimed at preserving and revitalizing the language. Built by Māori technologists, the software aims to ensure that Māori remains a living language in the face of increasing globalization and the dominance of English.

Challenges of AI

AI for Community also examines the challenges that can arise when cultural content is processed by AI tools that lack the right context. For example, facial recognition systems have shown inconsistencies across different populations, and image generators have occasionally misrepresented homes, traditions, or attire. These issues underscore the need for tools to be shaped by people who understand the nuances of the cultures involved.

AI for Community says AI holds promise for the future of language and linguistics. There are challenges to overcome, such as ChatGPT language outputs, which are described as ‘stilted’. In addition, the authors suggest there is a risk that revived languages become AI creations, not a continuation of the historical form.


Disclaimer: AAAS and EurekAlert! are not responsible for the accuracy of news releases posted to EurekAlert! by contributing institutions or for the use of any information through the EurekAlert system.