Skip to content

Jocotoco’s acoustic AI helps Kichwa communities protect 5,000 hectares

Jocotoco’s acoustic AI classified two million Amazon recordings in three months and gave two Kichwa communities species data with which Yakuwarmi protected more than 5,000 hectares in two areas.

  • 2 for an organization

    The foundation classified more than two million recordings in three months instead of a year or more.

  • 3 for society

    Two Kichwa communities used the species data in their management plans, and more than 5,000 hectares of their territory became Ecuador’s first community-declared conservation areas.

Depth runs from 1, a small gain, to 5, survival.

Status
Realized benefit.
Evidence

foundstudiedreported, the grade of this benefitclaimed

An independent news report supports the benefit.

Certainty

Very low.

Starts very low because no comparison was measured.

Size

0.28 =3+2+5+420×3+110

Scale3The highest depth above
Reach2Two Kichwa communities and one local organization used the data. Source
Duration5Two protected areas were created, a lasting change to the territory's status. Source
Distribution4The benefit reached two Indigenous Kichwa communities, which the source describes as the first local inhabitants in Ecuador to declare protected areas. Source
Probability3Realized.
Frequency1One species list led to one pair of declarations. Source
Compared with

Manual analysis of recordings from analog recorders. Assumed. No one measured it.

The system’s part
The AI classified the recordings and flagged threats, which produced the species list that the communities used. Jocotoco’s scientists retrained the models and validated the results, and the communities and Yakuwarmi made the conservation decisions. The figures come from Mongabay Latam’s reporting.
Limits
The figures come from one report of the foundation’s work, and the data from the latest processing were still awaiting validation. The report does not say when the two areas were declared.
Trade-offs
No costs or harms are reported in the sources.
Learns from data

Shown. Mongabay Latam reports that the foundation’s scientists retrain foundation AI models on large quantities of sound recordings.

“es reentrenar modelos de IA fundacionales” Source

Later outcome
Nothing later is known. The benefit was first reported on 28 May 2026, and no later report was found.
System
Fundación Jocotoco’s bioacoustic AI classifier (retrained foundation audio models)
Developer
Fundación Jocotoco (data science team)
Deployer
Fundación Jocotoco, with the Kichwa communities of Santa Rita and San Francisco and the organization Yakuwarmi
Where
Napo province, Ecuador (Narupa reserve, Cotundo, and the Kichwa communities of Santa Rita and San Francisco)
Stage
In a pilot
Date
, when the source was published
Kind
Benefit created
Broad benefit
Positive environmental impact, Social improvement

Sources

  1. News report, by authors with no tie to the system.

    Verdezoto Landívar, G., Comunidades kichwa lograron proteger su territorio gracias a la IA entrenada con miles de sonidos de la Amazonía ecuatoriana, Mongabay Latam, 28 May 2026.

    Yakuwarmi, una organización local liderada por mujeres, usó la información para conservar más de 5000 hectáreas de su territorio con la creación de dos áreas protegidas.
    La IA, que también capta sonidos de amenazas como motosierras y disparos, ha permitido a dos comunidades indígenas kichwa proteger y planificar su territorio.
    La Fundación Jocotoco recopiló más de 2 000 000 de audios entre 2023 y 2025. La IA les permitió clasificar y analizar esta vasta cantidad de datos en solo tres meses.
    Este encadenamiento de pasos fue clave para que los territorios de las comunidades se convirtieran en las dos primeras áreas protegidas por pobladores locales en Ecuador, proceso conocido como Otras Medidas Efectivas de Conservación Basadas en Áreas (OMEC).
    lista confirmada de especies que entregaron a las comunidades kichwa Santa Rita y San Francisco para que estas, a su vez, puedan utilizar los datos científicos en sus planes de manejo
    Los datos están esperando ser validados.

Details

AI system

Workflow

Foundation audio models retrained by the foundation’s data scientists on recordings from the Amazon. Recorders in the forest and, since October 2025, monitors that process sound in real time. Communities that put the species data into their management plans and an organization that carried the declaration through.

Follow-up

Conditions for replication

The recordings must be collected and the model retrained for the local species. The classifications must be validated by scientists before they are handed over. Communities need a legal route, here the OMEC process, to declare protected areas.

The entry

Sector

Biodiversity monitoring and community conservation. Non-profit conservation with Indigenous community partners.

Purpose

Fundación Jocotoco retrains acoustic AI models to identify the species and the threats recorded in Amazon forests, so that communities and the foundation can plan conservation.

Cite this benefit

Peter H. Vartanian. (2026). Jocotoco's acoustic AI helps Kichwa communities protect 5,000 hectares (AIBD-1239). Artificial Intelligence Benefits Database. https://w3id.org/aibd/AIBD-1239 (archived at https://doi.org/10.57967/hf/11470)

BibTeX Data What the fields mean