Digital Green Uses New Model for International Development

Digital Green's chatbot guides these farmers to information they need.

Not-for-profit firm applies artificial intelligence to support farmers

By Jake Smucker

Long-time MEDA supporters are familiar with MEDA’s unique model for international development. Digital Green, a small not-for-profit company, deploys an entirely new type of model — artificial intelligence large language models (LLMs). How can artificial intelligence (AI) play a role in the life-changing work of developing agricultural solutions to poverty?

Alesha Miller, Digital Green’s chief strategy officer, explains the organization’s specialty. “We are … co-creating a world where farmers use data and technology to build more prosperous communities,” Miller said. “That’s our mission. And we have reached six million farmers over the course of our 16-year history in 17 countries.” The funny thing about tech-centric organizations is that their business models change as new technology is developed. That is exactly what has steered the course of Digital Green’s history. With every twist and turn, Miller focuses on the problem: a lack of access to timely, quality agricultural advice for small-scale farmers.

Alesha Miller is Digital Green’s chief strategy officer. | Photo by Jake Smucker

Digital Green’s origins

A 2006 pilot project at Microsoft Research India (MRI) was driven by the idea that technology could help rural farmers. In 2007, Microsoft Research India decided to fold the project. However, a few of the researchers were inspired by the results they found. Rikin Gandhi and a couple of co-founders left MRI to continue the work with a new, independent organization. They named the organization Digital Green. Gandhi became its CEO.

Finding Success with Community Video

The first breakthrough technology for agricultural extension was surprisingly simple: video. Digital Green’s partners and trainers used miniature projectors to play instructional videos for rural Indians and Ethiopians. To their surprise, they often found that the whole community showed up. In some cases, it was one of the first videos the farmers had ever seen. By teaching through video, Digital Green increased farmer incomes by 24 percent.

The Global Pandemic

Digital Green reached farmers in 17 countries over the next decade. In 2020, it faced the same catastrophic interruption as the rest of the world — the COVID-19 pandemic. This forced the company’s next major change. Instead of hosting community video events, Digital Green started communicating virtually with farmers.

AI: The Future of Tech

Amid the pandemic, AI researchers made several breakthroughs in artificial intelligence. When ChatGPT launched publicly in 2022, the developers at Digital Green saw an opportunity. They started to dream about chatbots that could instantly educate farmers using regionally specific advice in the farmer’s native language.

Today, smartphones and access to video are found across most communities where Digital Green works. Though community video continues to be effective, it is no longer as revolutionary as it once was. Digital Green is finding that AI chatbots may hold the key to the future of agricultural education.

How Do Chatbots Help Digital Green Educate Farmers?

Through a popular messaging app called Telegram, farmers can ask questions of Digital Green’s chatbot, Farmer.Chat, in several languages: English, Hindi, Kishwahili, or Hausa. A chatbot will almost instantly reply with advice, tips, and relevant YouTube videos. During the chatbot’s testing phase, answers are checked by a human extension agent for accuracy and clarity. However, Digital Green sees a future where their chatbot is accurate and consistent enough that it would not need human oversight.

Early on, Digital Green relied on community video to provide low-cost agricultural education. | Photo courtesy of Digital Green

Cost-effective Agricultural Extension

From its start at MRI, this research sought to reach and educate more farmers using technology. Digital Green is finding success with this approach. Traditional, in-person face-to-face ag extension costs about $35 per practice that is adopted. That means it costs about $35 to teach a farmer a certain type of cover cropping or how to use a method of fertilization.

Digital Green found that community video reduced the cost of each practice adopted to $3.50. The organization expects to reduce the cost to 35 cents per practice using AI chatbots.

What can MEDA learn from Digital Green?

MEDA and Digital Green overlap in many areas of interest but do not cross paths directly — at least not yet. Miller can compare the organizations well because she works at Digital Green and hears stories from a family member who works at MEDA. “MEDA and Digital Green share a lot of the same goals, but we are working on different parts of the problem,” Miller identified. “I think that’s really important for development organizations to be very focused about where their particular special sauce is.”

Digital Green’s expertise in technology makes it unique in the international development world. MEDA is respected for its ability to build sustainable markets for small-scale farmers. Yet, both organizations focus on improving livelihoods for subsistence farmers — particularly women — in the Global South. “Agriculture is probably the best engine for poverty alleviation that the world has,” Miller said.

She believes that both MEDA and Digital Green are great at finding and supporting local leaders to help their entire communities out of poverty.

Digital Green’s Future

The rise of artificial intelligence gives reasons for excitement and concern. But both enthusiasts and skeptics of AI can celebrate the positive impact Digital Green is making using AI chatbots.

With every era and pivot in its history, Digital Green has made research central to its process. As technology evolves, the company continues to ask what’s next. Looking to the future, Miller said, “We continue to keep researching and experimenting with, ‘What are the other ways that technology can help solve these problems for farmers?’”


Latest Featured Stories

Digital Green Uses New Model for International Development

July 2, 2024

Digital Green uses artificial intelligence large language models to provide access to timely, quality agricultural advice for small-scale farmers.

Serving in Puerto Rico

July 2, 2024

Mennonite missionaries and entrepreneurs played a big role in developing hospitals and the poultry and dairy industries in the island...

Mowing a Path to Greener Solar Farms

July 2, 2024

Swap Robotics develops battery-powered machines that do large-scale grass and vegetation maintenance. The company wants to lead a worldwide move...