Wildlife traficking predictor

Unknown profit model, Unknown IP model, Ideation Phase, Open to new members
Thanks to algorithms and AI programs, cross-referencing data on deforestation and wildlife trafficking could help us anticipate wildlife trafficking incidents when deforestation occurs.

The Problem

As the world most biodiverse continent, Latin America is a hub for timber and wildlife trafficking. Each year, the criminal trade in wildlife generate billions of dollars for transnational criminal organizations. However, although drivers for wildlife trafficking are well known, the linkages with other criminal activities have been less documented. For instance, when illegal logging occurs in a given area, wildlife species are either captured or chased away. Our hypothesis is that, when deforestation occurs, wildlife trafficking is likely to increase nearby. Sometimes, this may just be a subsidiary activity. When they extract minerals from tropical forest regions, artisanal and small-scale miners get occasionally involved in wildlife trafficking. In other cases, illegal loggers cut all the trees except one, for animals to take refuge there, and so they can be captured more easily. However, while we assume that there exist points of convergences between deforestation and wildlife trafficking, the evidence provided is often anecdotal and does not elucidate possible systemic overlap of those two phenomena. Where and how these overlaps occur? If we could better understand the impact of deforestation on wildlife trafficking (in particular, the underlying dynamics / logics / causal relationships of such convergences), we believe we could better anticipate such traffics, and therefore combat them more effectively and efficiently.

Our Proposal

Our Wildlife Trafficking Predictor is a new app capable of identifying areas, animals and smuggling routes where wildlife trafficking is most likely to occur, after an area has just been deforested. Our proposed solution would therefore permit competent authorities in charge of combatting wildlife trafficking, to quicker and better decide where to intervene in priority, and to anticipate what animals would likely be smuggled and by whom. How to proceed: • Phase 1: Identify and map wildlife trafficking events and screen them against deforestation data. In some cases, correlations might be strong; in other cases, not at all. By cross-referencing wildlife trafficking incidents with satellite imagery of deforestation occurred in the same area a few days before, the idea is to identify points of convergences between deforestation and wildlife trafficking. • Phase 2: After identifying points of convergence, next step is to detect common characteristics that explain such linkages: where do such convergences most occur, how and why? The objective, here, is to acquire a better understanding of the dynamics between deforestation and wildlife trafficking and identify possible causal links, in order to create an analysis model. • Phase 3: based on this analysis model, next step will be to develop an algorithm that calculates the probability of wildlife trafficking in different contexts of deforestation, and then permit our system to raise red flags when this is most likely to occur.

We Assume that...

1- Convergences exist between deforestation and wildlife trafficking events, not just correlations and common characteristics/patterns exist beyond those convergences.

2- Wildlife trafficking incidents related with deforestation events occur depending on the environment (timber/wildlife species, other forms of transnational crimes).

3- It is possible to identify working patterns of such linkages.

4- Our new wildlife trafficking early warnings app would be cost effective and could be easily implemented.

5- No technical expertise is needed for the end users.

6- No effective or parallel system exists.

Constraints to Overcome

1- Access to and analysis of real time satellite imagery. Satellite imagery can detect deforestation activity in real time. However, access to real time images can sometimes result difficult because of poor weather conditions (clouds) and /or poor capacity to analyze those images. AI and machine learning programs can do the job automatically, but they are not yet available everywhere. 2- In specific regions, wildlife habitat database and timber species databases are often uncomplete or under construction and this could delay the functioning of our app. However, we believe that there are already enough data available to develop and test an analysis model and an algorithm for our early warning system.

Current Work

Our Red CORAL platform monitoring and mapping criminal activities in Latin America already has hundreds of data (especially for Colombia) and we will continue fueling our platform with new data every day, to obtain sufficient data regarding wildlife trafficking. We have also been discussing with Colombian authorities to evaluate the possibility to dispose of databases of forest species, deforestation and wildlife habitat. Next step we be to validate our hypothesis by cross-referencing sufficient data on deforestation with wild life incidents.

Current Needs

We need some funds and/or some AI experts to help us develop our analysis model and then the algorithm for our Wildlife Trafficking Predictor system. In a further stage, we believe machine learning programs could also help us automatically validate and improve our algorithm/early warning system.

or
 
DISCUSSIONS & TASKS
IN AREA / PROJECT