How do you protect a forest when you cannot be there to see what is happening? For Nigeria's forests, where illegal logging continues to threaten ecosystems and biodiversity, the answer may lie in technology that can detect human activity before the damage becomes more extensive.
On October 2, 2026, I interviewed Lesley John Jumbo, a robotics and embedded systems engineer and co-founder of Reforest AI, during Environmentalist View's Change Makers webinar series. Our conversation explored how his team is developing an AI-powered forest monitoring system designed to detect illegal logging and alert responders to potential threats. We also discussed the challenges of building environmental technology in Nigeria, the limitations of existing forest monitoring methods, and what it takes to turn an innovative idea into a solution that works in the field.
As Lesley explained during the interview, the idea behind Reforest AI is straightforward: give forests a way to signal when something is wrong, even when nobody is watching. However, building a system that can do this reliably requires more than promising technology. It demands practical engineering, real-world testing, and a workable plan for responding when a threat is detected.
An Idea Inspired by the Amazon
During our conversation, Lesley traced the origins of Reforest AI to a hackathon project and a question that emerged after he read about businessman Johan Eliasch's purchase of a section of the Amazon rainforest in an effort to protect it. The story made him consider the scale of the challenge. If protecting forests depended on purchasing land to keep it away from loggers, how much forest could realistically be saved that way?
He began examining existing forest monitoring methods and identified gaps he believed technology could address. Satellite imagery can help identify deforestation, but it may reveal damage only after it has occurred. Drones offer another means of monitoring remote areas, although their power requirements can make continuous surveillance difficult. What he wanted was a system that could remain on the ground, listen for suspicious activity, and alert someone while an incident was unfolding.
That became the foundation for Reforest AI. Rather than relying entirely on observing a forest from a distance, the team wanted to develop a monitoring system capable of recognising potential threats as they happened.
The idea raises an important question for forest conservation in Nigeria: could technology help close the gap between when illegal activity begins and when those responsible for protecting the forest become aware of it?
Why the Team Had to Rethink Its Design
One of the interesting parts of my conversation with Lesley was learning how the team had to reconsider its original design. The first concept involved placing a monitoring device on every tree. While the idea offered a way to monitor individual locations, it quickly became apparent that the approach would be difficult to scale.
Lesley recalled how a stranger at the Global Entrepreneurship Festival in Ondo State pointed out the problem. With billions of trees across the world's forests, deploying and maintaining a device on each one would be impractical. The team subsequently shifted towards using poles capable of monitoring a wider area.
The change illustrates a fundamental challenge in environmental innovation. A technology can be technically interesting and still fail to make a meaningful difference if it cannot be deployed at the scale the problem demands.
Lesley described the system as having ears and eyes. Microphones continuously listen for sounds associated with human activity, including chainsaws and axes. Machine-learning software classifies the sounds to determine whether they could indicate a threat. When a relevant sound is detected, cameras activate to provide additional confirmation.
The system is designed around three poles: two listening poles and a central unit the team calls the "hero" pole. The listening poles communicate detections to the central unit using LoRa, a long-range, low-power radio communication technology. The hero pole then sends an alert to a dashboard through a small satellite communication device.
Connectivity is a major consideration because forests often lack reliable internet access. Instead of attempting to transmit large audio and video files, the system sends a short text alert containing a timestamp and GPS coordinates. The recorded audio and video remain stored on an SD card within the device.
This approach reduces the amount of information that must be transmitted while giving responders an indication of where a possible incident is occurring. It also highlights an important distinction: the system is designed to detect and document suspicious activity, not to independently apprehend offenders or guarantee that illegal logging will stop.
Building With Limited Resources
Developing environmental technology in Nigeria presents challenges that extend beyond the engineering itself. Equipment that might be readily available elsewhere can be difficult to obtain locally, and importing specialised components is not always an option.
Reforest AI encountered this problem when the team could not obtain a listening module it wanted to use. During our interview, Lesley explained how they adapted rather than abandoned the design. The team purchased inexpensive microphones, and he developed sound-classification software that could run directly on the device.
The prototypes are manufactured using 3D printing, with PLA and PETG plastics used for the casings. An aluminium pole is planned for the final version, while the current design incorporates forest-green and brown colours to help the equipment blend into its surroundings.
These decisions reflect the practical compromises involved in building a system that must eventually operate outdoors, potentially for extended periods, without constant access to maintenance or technical support. The team has had to balance cost, component availability, power requirements, and the need for a design that can withstand the conditions in which it will be used.
For Lesley, these constraints have also encouraged resourcefulness. The challenge has been to work with what is available while continuing to develop a system that can meet the demands of the environment it is intended to protect.
This is particularly relevant for Nigerian innovators working on environmental problems. Limited access to equipment and funding can slow development, but building around local realities may also help produce solutions better suited to the environments in which they will eventually operate.
Detecting Illegal Logging Is Only Half the Problem
One of the most important issues that emerged during my interview with Lesley was the distinction between detecting illegal logging and actually stopping it.
A monitoring system can identify suspicious activity, record evidence, and alert the appropriate people. But what happens next?
Lesley acknowledged that while the team has made progress on detection, the enforcement side of the system remains a work in progress. Identifying a possible logging operation does not automatically mean that anyone will respond, arrive in time, or be able to intervene safely.
Illegal loggers may be armed, making direct confrontation dangerous for forest rangers and other responders. Any enforcement model must therefore consider not only how quickly an alert is delivered but also who receives it, what authority they have, and how their safety will be protected.
The team's proposed initial approach includes placing speakers away from the monitoring poles. These speakers would play warning sounds intended to deter intruders without revealing the precise location of the equipment. The objective is to discourage the activity while responders are being mobilised.
However, the effectiveness of this approach will depend on how people react to the warnings, whether potential offenders can identify the source, and how quickly a response can be organised. A warning system may discourage some intruders, but it cannot be assumed to deter determined or armed offenders in every situation.
Lesley also highlighted the potential role of government in training and protecting responders and facilitating access to restricted forest areas for testing. At the same time, he acknowledged a difficult institutional reality: some officials may themselves be connected to the activities that forest protection efforts are intended to prevent.
That tension makes enforcement more complicated than installing sensors and connecting them to a dashboard. Technology may improve detection, but effective forest protection also requires trustworthy institutions, clear responsibilities, and people who can act on the information received.
It was a reminder from our conversation that environmental problems rarely have purely technical solutions. The technology can help identify the problem, but the institutions and people responsible for acting on that information remain just as important.
Why Field Testing Matters
Like many emerging technologies, Reforest AI faces a gap between what works under controlled conditions and what happens in a real forest. A system that correctly identifies chainsaw sounds in a laboratory may behave differently when exposed to wind, rain, insects, birds, machinery, and other sounds commonly found in forest environments.
Lesley explained that the team expects these differences to affect detection accuracy, including the possibility of false negatives, where actual threats go undetected. False positives are another practical concern because harmless sounds could trigger alerts and waste the time and resources of responders.
Rather than assume that laboratory performance will translate directly into reliable field performance, the team plans to test the technology in forests, collect data, and refine the system based on what it learns.
This process will be essential to establishing whether the equipment can operate consistently under real conditions. It will also help the team understand the practical limitations of its approach, from sound classification and communications to power supply, weather exposure, and the reliability of its alerts.
For Reforest AI, the next stage is not simply to demonstrate that the technology works. It is to establish where it works, where it fails, and what must change before it can be deployed more widely.
Nearly Giving Up, Then Finding Recognition
During the webinar, Lesley also spoke about a period when the team came close to abandoning the project. A mentor had told them that investors did not care about deforestation, a discouraging assessment for a team attempting to build a business around an environmental problem.
The breakthrough came through the 2025 National Geographic Society Slingshot Challenge, where Reforest AI won top honours and received a US$10,000 grant. The recognition provided financial support and an important signal that the idea could attract attention beyond the team itself.
However, the experience also taught Lesley a lesson about how environmental solutions must be presented to potential buyers.
The team had initially assumed that environmental agencies and non-governmental organisations would be willing to purchase the technology because of its environmental benefits. That assumption proved too simplistic. Organisations may support forest conservation as a goal, but purchasing decisions also depend on budgets, operational needs, measurable outcomes, and the risks a solution can reduce.
A prospective buyer needs to understand what the system offers in practical terms. Can it reduce the cost of monitoring a forest? Can it help prevent losses? Can it improve response times or provide useful evidence when illegal activity occurs? These are the kinds of questions that determine whether a promising environmental innovation can become a sustainable business.
This is an important lesson for environmental entrepreneurs. Having a solution to a genuine environmental problem does not automatically create a market for it. The environmental benefits must be matched by a clear explanation of the value the solution provides to the people and organisations expected to pay for it.
For Reforest AI, recognition was an encouraging milestone, but the next challenge is to demonstrate that the technology can deliver results that justify investment.
Beyond Catching Illegal Loggers
Our conversation also explored how Reforest AI's technology could potentially be used for purposes beyond detecting illegal logging.
Lesley discussed the possibility of using the system's stored recordings and camera footage to support timber verification. Under the European Union's deforestation regulation, businesses dealing in covered commodities, including wood, face requirements intended to prevent products associated with deforestation and forest degradation from entering the EU market. This creates a need for credible information about where timber originates and whether it meets the applicable requirements.
If Reforest AI can produce reliable, time-stamped records linked to specific locations, those records could potentially contribute to monitoring and verification processes. They might help document activity in monitored forests and support investigations into suspected illegal logging.
However, there is an important qualification. Recordings do not automatically establish the legal origin of timber, and their evidentiary value will depend on data integrity, the ability to verify where and when recordings were made, and whether regulators and buyers accept the information for the purpose in question. The technology would need to operate alongside appropriate traceability and verification procedures rather than replace them.
During the webinar, a listener also raised the possibility of adapting the technology to help farmers protect their land and detect theft. Lesley acknowledged the idea as another potential application the team could explore.
These possibilities suggest that the underlying technology could have uses beyond its original purpose. Nevertheless, each application would require its own testing, validation, and assessment of practical value. A system designed to identify sounds associated with illegal logging, for instance, would need to be evaluated separately before being relied upon to detect theft or other forms of unauthorised activity.
The broader opportunity is interesting, but the priority remains proving that the original application can work reliably in the field.
The Calabar Test: From Prototype to Forest
The next significant milestone for Reforest AI is a planned deployment in a forest near Calabar in November 2026. During our interview, Lesley explained that the deployment follows a request from people in the area and is expected to provide an opportunity to test the system under real forest conditions.
The team also plans to produce a documentary and an explainer video around the work. These materials could help communicate how the technology operates and document the lessons emerging from the deployment.
For now, the Calabar test represents an important step towards determining whether Reforest AI can move from a promising prototype to a dependable forest monitoring solution. Its performance will need to be assessed against practical questions: whether the microphones consistently detect relevant sounds, whether alerts reach the dashboard, whether the recorded evidence is useful, and whether the system can function reliably in the conditions for which it was designed.
Lesley has set himself and the team an ambitious five-year target. He wants verifiable evidence that Reforest AI has helped protect more than ten forests from specific incidents of illegal logging, supported by identifiable forest locations, dates, times, and relevant data.
It is a measurable target, and that is what makes it important. Counting installations alone would show that equipment had been deployed, but it would not establish whether the technology had made a meaningful difference. Evidence of specific incidents prevented or addressed would offer a stronger basis for evaluating its impact.
The planned deployment near Calabar will therefore be about more than testing hardware. It will be an opportunity to gather evidence about the system's practical capabilities, identify its weaknesses, and understand what would be required to expand its use.
Until those results are available, Reforest AI's potential should be treated as a promising possibility rather than an established outcome. The field test will help determine how much of that promise can be translated into measurable results.
Lesley's Advice to Young Nigerian Innovators
Towards the end of our conversation, Lesley offered advice to young Nigerians who want to develop solutions to environmental and other societal problems. His argument was that advocacy matters, but action matters too. Raising awareness about a problem is important, yet building something that can address it offers another way to contribute.
For those without funding or technical skills, he advised sharing ideas online, using free tools, and starting small. He pointed out that Reforest AI itself gained attention through social media, demonstrating how making an idea visible can create opportunities to connect with people who may be able to help.
His experience also offers a useful reminder that innovation does not always begin with access to the best equipment or substantial funding. Sometimes, progress comes from identifying a problem, building a workable first version, recognising its weaknesses, and improving it through testing.
At the same time, young innovators must be prepared to question their own assumptions. An idea that appears useful to its creator may not immediately appeal to potential customers. A prototype that works in a controlled environment may struggle outdoors. And a system that detects a problem may still need partnerships, institutions, and operational plans to produce the desired outcome.
These are not reasons to abandon innovation. They are reasons to approach it with curiosity, discipline, and a willingness to learn from evidence.
The Work Ahead for Reforest AI
My conversation with Lesley John Jumbo offered a look at both the possibilities and the difficulties involved in developing technology for environmental protection in Nigeria. Reforest AI is attempting to address a genuine challenge by combining sound detection, machine learning, cameras, low-power communication, and satellite connectivity in a system designed for remote forests.
The planned Calabar deployment marks an important next stage for Reforest AI, offering an opportunity to test and refine the technology in real forest conditions. What stood out to me during the Change Makers interview was the ambition behind the project: using AI to help make illegal logging easier to detect and forest protection more responsive.
As the team moves towards deployment, its work could contribute to new approaches to forest monitoring in Nigeria. Ultimately, the goal is not simply to hear what happens in the forest, but to turn those sounds into useful information that supports timely action and helps protect the ecosystems on which communities and wildlife depend.
Comments
Post a Comment