Urgent call to protect Madagascar’s biodiversity through AI

A large-scale analysis of Madagascar’s unique biodiversity has just been conducted with the participation of over 50 international organizations. Madagascar is one of the world’s biodiversity hotspots, with 82% of its plant species and 90% of its vertebrates found nowhere else on Earth. The results of the study, which rely on AI-based tools developed by Daniele Silvestro of the Swiss Institute of Bioinformatics and his team at the University of Fribourg, were published in two papers in the journal Science. They reveal an urgent need for science-based collaborative conservation measures. The researchers formulate proposals for future actions in this direction.

Combining recent resources and state-of-the-art techniques to predict species conservation status, the international team assessed threats to terrestrial and freshwater biodiversity, and then examined future conservation and restoration opportunities.

Daniele Silvestro, group leader of the Swiss Institute of Bioinformatics (SIB), and his team contributed by predicting the extinction risk and potential threats for 6000 Malagasy plant species. Assessing the conservation status of species is the first step in developing effective conservation measures. The problem is that for the majority of organisms, especially among plants, fungi and invertebrates, such an assessment is still lacking. Thanks to the methods developed, we were able to predict the extinction risk of several thousand plant species, by analyzing available online databases with machine learning algorithms’, says Daniele Silvestro. To this end, the scientists have developed an open source software based on a form of artificial intelligence called ’Bayesian neural networks’. The idea, outlined earlier this year, was to develop open source software using artificial intelligence to better identify regions and species in need of protection.

Analyze to better protect
The results of this study highlight Madagascar as one of the highest priority regions for conservation. The researchers gained in-depth knowledge of the island’s past and present diversity, as well as its current distribution and future threats. They conclude by calling for a sustained acceleration of data collection and analysis to preserve Madagascar’s unique diversity. They make proposals for future fair and equitable conservation measures, ranging from increased data generation and availability to addressing the root causes of biodiversity loss, including poverty and food insecurity.

> Ralimanana H. et al, Madagascar’s extraordinary biodiversity: Threats and opportunities , Science
> Antonelli A. et al, Madagascar’s extraordinary biodiversity: Evolution, distribution, and use , Science
> Zizka A. et al, IUCN - Deep learning approaches to approximate species’ extinction risk. Diversity and Distributions , Diversity and Distributions