For most of my career, I’ve tried to ensure that people have the information they need to make decisions that safeguard nature. Yet the people whose choices often matter most for forests, fisheries, and wildlife are usually the least served by conventional media.
Farmers, fishers, Indigenous land stewards, and community leaders often operate in what I’ve come to think of as “information deserts”: critical insights exist, but rarely reach the frontlines where ecosystems either endure or fall apart.
This is the gap we are closing with StoryTransformer, a tool that will enable us to adapt trusted environmental reporting and scientific knowledge into forms that are useful to local decision-makers whose choices determine the fate of habitats on a micro level. Our focus isn’t audiences who…
For most of my career, I’ve tried to ensure that people have the information they need to make decisions that safeguard nature. Yet the people whose choices often matter most for forests, fisheries, and wildlife are usually the least served by conventional media.
Farmers, fishers, Indigenous land stewards, and community leaders often operate in what I’ve come to think of as “information deserts”: critical insights exist, but rarely reach the frontlines where ecosystems either endure or fall apart.
This is the gap we are closing with StoryTransformer, a tool that will enable us to adapt trusted environmental reporting and scientific knowledge into forms that are useful to local decision-makers whose choices determine the fate of habitats on a micro level. Our focus isn’t audiences who scroll polished websites; it’s the herder deciding whether to plant trees, the farmer weighing whether to clear mangroves, the fishing cooperative debating a seasonal closure, or the village leader under pressure to sell community forest land.
For many, artificial intelligence can feel like part of the problem: extractive, inequitable, resource-intensive. Those concerns deserve to be taken seriously. That’s why we are building safeguards into the system from the start and making sure human judgment remains central.
Here are four key principles guiding StoryTransformer:
- This doesn’t replace journalists. This will be expansionary, enabling us to hire more editors and native-language contributors, not fewer. AI will assist with adaptation, but every output will still go through humans who ensure accuracy, context, and cultural and topical relevance. Nothing publishes without people in the loop.
- We are designing this responsibly. That means grounding content in verified journalism and science, double-checking model outputs, and building in a strong review process. No hallucinations, no synthetic “invented” stories; only adaptations of work we’ve already rigorously reported
- We are thinking about our environmental footprint. From efficient query construction to lower-power inference, we’re working to ensure impact does not come at the expense of the planet. We do not want to solve one environmental problem while contributing to another.
- We are strengthening information ecosystems. By hiring native-language editors, we support a two-way flow: not just distributing knowledge outward, but elevating local priorities and perspectives back into global awareness.
In practical terms, imagine:
- A story on the livelihood benefits of a seasonal fisheries closure turned into a 90-second WhatsApp audio message in Nigerian Pidgin
- A piece in Tagalog for community radio on how a community won recognition of its territorial rights.
- Tips on effective approaches for tree planting delivered as simple SMS updates in Marathi.
This is not automation for its own sake. It’s about accuracy, agency, and relevance — not scale for scale’s sake.
Where we are now: We’ve spent 15 months laying the groundwork and building a prototype. Dedicated funding now allows us to accelerate development. The plan is to release a working system in six languages by the end of 2026, starting with regions where Mongabay already reports deeply and where measurable audience needs are strongest.
Longer-term, we aim to support Indigenous languages that are primarily spoken — not widely represented on the Internet — while exploring collaborations with organizations developing specialized models for under-served languages.
Formats will reflect how people actually stay informed: short audio messages for WhatsApp and community radio, text for SMS, and eventually a chat interface that answers questions using Mongabay’s journalism and can say “I don’t know” when it doesn’t have an answer.
What StoryTransformer will not do: ✔ Generate original articles ✔ Publish without editorial approval ✔ Replace reporters or editors in the languages we currently publish in
And we will measure impact not by clicks, but by whether people feel better equipped to make decisions that protect their environment and livelihoods.
I’m thrilled to share that an individual and the OpenAI Foundation have provided seed support for StoryTransformer, a first step toward realizing this vision.
If we succeed, this will help shape everyday choices in places where nature’s future is being decided and create new livelihoods for the people who live and work there.
Life on Earth deserves nothing less.