Indigenous Youth Are Using Coding and AI to Save Native Language

A wide grin spread across Niesha Marshall’s face as she watched the purple go-kart zip across her computer screen. It worked. The game worked. “I was really proud of myself…I just couldn’t believe it, that I created this AI,” she said.

It was the summer of 2021 and Niesha, then 13, moved to campus just three weeks earlier knowing almost nothing about AI. She hadn’t been to school in over a year because of COVID and she just wanted to get out of the house, to see someone other than her family. But all she knew of coding was what she’d seen in movies: a geeky guy alone in the corner hacking into the government, or something like that, she recalled.

In less than a month, her view on AI and who could participate in building it radically shifted.

Niesha is one of the high school-aged students to complete the Lakota AI Code Camp — a three week summer intensive taught by four Indigenous AI experts from across the country. At Black Hills State University, about 220 miles northwest of the Rosebud Reservation, the camp trains Lakota students with little coding experience in Python, data science, machine learning, and app development in less than a month. It’s a foray into computer science for interested young people, but the camp is also about more than creating opportunities for individual students. It’s an attempt to build an Indigenous talent pool, to find and train future experts who can digitally protect and steward Indigenous culture — and keep endangered native languages alive.

Indigenous languages are at risk. According to the United Nations, at least half of all the world’s languages will disappear by the end of the century, the majority of which are Indigenous languages. In the U.S., assimilation policies rooted in centuries of racism play a large part in endangering native languages and culture. For generations, Native American children were punished for practicing their native language and customs. Without intervention, these languages — and the cultural knowledge embedded in them — could be lost within a generation.

AI offers hope for revitalization, a way to fortify and teach languages, to record traditional knowledge, and to facilitate customs virtually. That’s the sort of work the camp co-founders, Mason Grimshaw and Michael Running Wolf, are hoping to achieve.

But they both ran into the same problem: They needed more help and the workforce of Indigenous AI and machine learning experts was in extremely short supply.

Research has shown that 18% of those who live on tribal reservations don’t have internet access. And, while Native Americans make up roughly 2% of the U.S. population, they only account for a fraction of a percent of computer science faculty — and Running Wolf says there don’t seem to be many more professionals coming up the ranks. Native students earn only 0.1 percent of doctoral degrees in computing, according to a 2023 report. “We have a problem,” Running Wolf, an AI ethicist and language revitalization expert who is also Lakota, said.

“My interest is to uplift communities and make sure they can be owners of this technology and have agency over the AI,” Running Wolf said. “But no one is going to build that for us.”

So Grimshaw, a data scientist who is Lakota, suggested they build a talent pathway. Not enough native students are choosing computer science, so why not try to reach high school students before they make those decisions? “[Native students] say no to so many things implicitly because they never see it,” Grimshaw told Teen Vogue, stressing the importance of representation for Indigenous people in computer sciences.

At first the founders didn’t think the camp would succeed because the professional educators they consulted all had the same response: there’s no way to teach a 13-year-old Python if they’ve never coded before. That’s not even to mention the technical difficulties the group faced. Typically, machine learning can be done in the cloud, but the internet in rural South Dakota was unreliable according to Grimshaw. Instead, every student at the camp needed a gaming computer so machine learning technology could be run on the device. And still, a routine Windows update of the seven computers brought the internet down across the entire university, Running Wolf said.

But they tried anyway just to see how far the kids could go, and they went far. “We ran out of content, finished a week’s worth of content, by day three,” Running Wolf said.

There were still moments of self-doubt among the students, Grimshaw said. One student slipped out to call her parents and ask them to come get her. She felt like the only one who couldn’t get it, he recalled. But with a little extra help, she was back on track. Encouragement from people who can understand their background, Grimshaw said, is invaluable. “I can say, ‘We’re from the same reservation. And I did it. So, you can do it too.’” For Niesha, that inherent understanding helped her push through struggles at camp. “One of the best parts of the camp, you can use native humor and people would understand,” Niesha said. By the end “the camp felt like family to me.” In fact, one of the reasons she plans to go back for a third summer is because of how the camp emphasizes her culture.

To integrate coding skills in the community and change the way natives interact with computing, “you need Indigenous people teaching it,” said Brendan David-John, PhD, professor of computer science at Virginia Tech and member of the Seneca tribe, who was not involved with the camps. That’s what makes expanding the camps challenging. Each one needs local teachers and tribe-specific curriculum, he said.

Tribe-specific curriculum is what can really push forward the crucial revitalization of culture. For their final project, the students developed an app that could recognize plants sacred to the Lakota. They took pictures of local flora and fauna during a hike in the Black Hills, sacred Lakota lands. Then with the help of a local knowledge keeper, an ethnobotanist who knew the stories and uses behind each plant, they labeled the data and trained their model.

There are plans for replicating the camps: one in California and David-John has been discussing the possibility of coding camp for Seneca youth in upstate New York. But the Lakota team won’t have the time or cultural expertise to run other camps. They do plan to train more local teachers, though. Grimshaw and Running Wolf are building a program now that will hopefully equip local STEM teachers and community leaders to start their own coding camps.