All ears listening to biodiversity
From radars to rainforests, AI has also found application in the depths of some of Southeast Asia's oldest landscapes.
Students Klara Hernblom and Johan Norway from the Swedish University of Agricultural Sciences are conducting research in the ancient – and degraded – forests of Sabah on the Malaysian island of Borneo.
They are the target To better understand levels of biodiversity and wildlife activity in different landscapes, including restoration sites. The findings may provide insights into the effectiveness of carbon credits, where companies can offset their carbon footprint by restoring or conserving forests.
To understand the environment, they placed dozens of audio recorders throughout the landscapes they studied. Over 10 days, the devices capture the sounds of the animals, birds and reptiles that live there.
Instead of manually listening to thousands of hours of recordings and trying to determine species and behaviors, students can upload their recordings to an AI-powered platform called Arbimon, which can analyze them in real time.
„It looks really promising. Without it this project would not have been possible. And going forward, we will only get more species that we can identify,” said Mr Narva.
It is a method widely used in the region and around the world – the online tool is free for scientists and researchers to access.
Orbemon was originally launched as a cloud-based project to store and analyze audio recordings at a university in Puerto Rico. In early applications, it helped park rangers detect illegal activities such as chainsaw use in the jungles of Palawan in the Philippines.
Its capabilities are now more complex.
The technology uses pattern matching and clustering to find specific sound signatures that indicate living organisms. It has the potential to provide insights into what is happening in the forest – in real time.
„For a scientist without access to technology, it takes an average of 10 to 15 minutes to process a single record,” said Mr Bourhan Yassin, CEO of Rainforest Connection, which runs Arbimon. The non-profit organization is dedicated to protecting threatened ecosystems.
„So think about that and multiply that by 100,000 records, which isn't much for a research paper or a thesis. It equates to somewhere around four and a half months to get really good detections for a species.
“With an AI model, you can do this in seconds. In fact, you can process a million records in seconds,” he said.
Mr Yasin said Arbimon helps bridge a huge gap between science and research and conservation on the ground. He shared that two to three million posts are uploaded to the platform every week. Over two billion analyzes have been performed so far in nearly 6,000 projects in nearly 120 different countries.
Despite the vast potential for the use of environmental acoustics, particularly in Malaysia and Indonesia, where rainforests are dense, there are limitations surrounding its use in Southeast Asia.
Mr Yassin acknowledged that governments are still wary of collecting data, particularly on national parks, and sharing that data outside their own countries.
„Government regulations haven't caught up yet,” he said. “For AI to be widely adopted in Southeast Asia, there needs to be decentralization and an understanding of the fact that AI does not have to be completely localized and cannot be completely controlled within the country.
„These systems, especially systems like Urbemon, don't just serve one country, they have to be global.”