Despite the persistent rain, the second day of Media Party New York 2026 began with an early surge of participants ready to dive into a marathon of global perspectives and transformative hands-on learning.
Opening the day’s sessions, Jacopo Ottaviani, Chief Data Officer at Code for Africa, introduced Africa Mining Watch—a project utilizing artificial intelligence to expose illegal mining across the African continent. His presentation focused not only on the technical process of satellite monitoring but on the essential human infrastructure required to turn data into long-term impact.
The Resource Curse: A Data Challenge
Ottaviani began by highlighting the “Paradox of Plenty.” While Africa holds an estimated 13% of the world’s mineral reserves, this wealth often correlates with lower economic growth and conflict. In the Democratic Republic of the Congo (DRC), the demand for resources —mostly those used in global electronics—fuels labor exploitation and corruption. For journalists, the sheer scale of the terrain has historically made monitoring these remote, often illegal, operations an impossible task.
From the Amazon to the Congo
The inspiration for the project came from across the Atlantic. Amazon Mining Watch, developed by journalist Joseph Poliszuk and the nonprofit Earth Genome, used machine learning to scan satellite imagery for illegal gold mines in the rainforest.
Why is this only working in the Amazon and not in Africa?” Ottaviani asked.
This sparked a consortium between Code for Africa, the Pulitzer Center, and Earth Genome to adapt the algorithm to the African landscape, resulting in a specialized instance of the technology tailored to the continent’s unique geography.
Earth Index: No-Code AI for Journalists
The engine behind this initiative is Earth Index, a no-code tool designed to democratize satellite data. The system follows a four-phase workflow:
- Ingestion: Processing massive volumes of satellite imagery.
- Detection: Identifying patterns consistent with industrial or artisanal mining.
- Scoring: Assigning a confidence level to each flag.
- Verification: Allowing humans to filter out false positives.
By removing the technical barrier of coding, the platform enables journalists to focus on investigation rather than software development.
The "Human-in-the-Loop" Philosophy
Ottaviani emphasized that technology is only half the solution. On Earth Day 2026, the project hosted its first Mappathon, where 25 African journalists spent seven hours training the AI model.
We need humans to humanize the data and contextualize the data journalism,” Ottaviani asserted.
This “human-in-the-loop” approach ensures that the AI isn’t a “black box” but a tool shaped by local expertise. By involving reporters in the training phase, the project fosters community ownership, turning digital coordinates into impactful stories about power and exploitation.
Breaking the Cycle: Sustainability through Ownership
Addressing a systemic flaw in the civic tech space, Ottaviani noted that many projects “start and end with a grant.” To counter this, Africa Mining Watch was designed for long-term sustainability by prioritizing journalist ownership over trainer dependency.
We wanted a tool that would become sustainable over time and used by the journalists who would basically take ownership of that,” he explained.
By utilizing an accessible interface and establishing one-on-one mentorship tracks, the project moves beyond a “one-off” training model. This approach builds a trust relationship with a selection of journalists who are not only equipped to use the tool but are also empowered to train their peers, ensuring the initiative remains resilient and community-driven.
The Path Ahead
To ensure the project survives beyond the initial phase, Africa Mining Watch integrates this community focus with access to investigative grants via the Pulitzer Center and a syndication network to amplify local stories globally.
The team plans to expand the project from the Kivu region to West and South Africa. An interactive map is set to launch in the coming weeks, providing a public resource to track illegal mining in real-time. For the Media Party community, Ottaviani’s keynote served as a powerful proof of concept: when AI is placed in the hands of a dedicated community, the invisible becomes visible.
Learn more and follow the project at Africa Mining Watch.
Stay tuned for more updates and deep dives from Media Party New York 2026!
