As the march of technology continues to press forward, it’s not just humans that have been benefiting from its advancement – researchers are finding ways to use new tools to help many vulnerable animal species as well.
Scientists from the Benioff Ocean Science Laboratory collaborated with the Marine Mammal Center to deploy technology-packed buoys off the coast of San Francisco as part of the “Whale Safe” initiative, intended to combat incidents of whale deaths due to collisions with boats and shipping freighters at sea.
“Whale mortality by ship collision is one of the top sources of endangered whale death, there’s nothing more tragic than seeing one of these whale-ship collision victims.
“About 80 endangered whales in the west coast of the US are killed every year, and that’s way too many whales, but this is a solvable problem,” said Douglas McCauley, the director of the Benioff Ocean Science Laboratory in a video posted on their YouTube channel.
The programme uses artificial intelligence-enabled (AI) acoustic monitoring instruments to detect the presence of whales in offshore waters, with buoys equipped with hydrophones – underwater microphones that detect whale vocalisations – positioned on the ocean floor, which sends the captured data to a small computer in the buoy and then via satellite to scientists for review.
These vocalisations can then be used to identify the species of whale and, when analysed by the AI algorithm alongside other data points like ocean conditions, recent sightings from trained naturalists, and location data from 104 satellite-tagged blue whales, can help predict the presence of whales in the area and the likelihood of an encounter.
Whale Safe then summarises this information with data from the Automatic Identification System (AIS) – a tracking system used by large ships and companies to navigate and avoid collision with each other – to make recommendations for vessels in shipping lanes that are likely to cross paths with a whale to adjust their speed, so as to minimise the risk of impact.
More AI-powered conservation efforts
In other parts of the world, conservationists have been using similar AI-powered technology to help protect different endangered species.
Machine learning systems have been trained to spot poachers in places like Zambia’s Kafue national park to stop illegal fishing, with thermal cameras set up along a 19km stretch across Lake Itezhi-Tezhi to identify illegal boat crossings and alert park rangers, removing the need for constant surveillance of the area.
Some initiatives use AI systems to observe changes in the environment – like Brazil’s loss of surface water – and others use it to manage the huge troves of data captured by trap cameras and microphones of endangered animals to identify and keep track of their populations.
Efforts towards conservation by Google alongside partners like the World Wildlife Fund (WWF) and Conservation International have also involved AI in this way, with their Wildlife Insights system that leverages photos taken from various camera trap projects around the world to understand how the populations of endangered species are changing.
“Camera traps have become an essential tool for studying wildlife. Often deployed in remote areas for long stretches, they can snap thousands of photos of animals that researchers rarely see up close.
“But sifting through all that imagery can take weeks, even months. Tagging and analysing the photos requires extensive training, and uploading and sharing them is a file-transfer nightmare,” the WWF wrote on their website, adding that deploying Wildlife Insights would leverage AI to speed up the process, and allow conservationists to act faster to protect wildlife.
The Wildlife Insights system has also previously been used to tag and analyse images taken during a mammal and bird monitoring programme at the Pasoh Forest Reserve in Negri Sembilan.
Initiatives like the Protection Assistant for Wildlife Security (PAWS) developed at the University of Southern California – to protect endangered animals like elephants, tigers, antelopes, deer, macaques and leopards, to name but a few – have also seen successful tests in Uganda and Cambodia from as far back as 2014.
Using a machine learning algorithm, the PAWS system takes data from previous poaching activities and uses it to predict where poachers are likely to appear.
These predictions are then used by the algorithm to suggest patrol routes for at-risk locations, which are also changed after a period of time so as to be randomised, preventing poachers from learning patrol patterns.
More recent developments in the use of AI in wildlife conservation efforts tend to include drone technology as well, such as the SnotBot – a drone equipped with a camera and a petri dish – that’s meant to capture whale photographs and snot samples containing DNA, which can give scientists valuable data on whales.
Algorithms used to identify poachers and survey the population of vulnerable animal species have also been combined with these camera-equipped drones.
This includes the efforts being made by Madeline Hayes, a drone pilot and graduate student at the Boston University in Massachusetts, with her AI model trained to identify and count both rockhopper penguins and black-browed albatrosses on the Falklands Islands in photos captured by drones, as well as projects like Conservation AI that combines poacher-identifying machine algorithms along with images taken by drones.