A lot is written about the commercial benefits of integrating AI and using data to generate value. There’s no doubt that both can make a big difference to the bottom line of any business. But that’s not where the opportunities in this innovation end. We are increasingly seeing AI being used to make a positive impact in the world, tackling crises that have been unsolved for decades and bringing new light and hope to challenging circumstances. Wildlife conservation is one of these crises – AI has now been identified as one of the top 3 emerging technologies in conservation, an effective force for good.
The escalating biodiversity crisis
Biodiversity is the myriad of all living things on earth and how they fit together. Crisis mode has been activated because scientists have warned that we are living in a sixth age of extinction. In a recent UN report, a panel of intergovernmental scientists identified that more than a million plant and animal species now face being wiped off the face of the earth. According to the WWF’s latest Living Planet Report, between 1970 and 2016 the global populations of mammals, birds, reptiles, amphibians and fish plunged by 68%. It’s clear that humans – and our innovations – are often not good for the natural world. But could we be better?
AI as a force for good
Existing approaches to wildlife conservation have proved to be frequently ineffective. For example, traditional methods of surveillance and monitoring can be inefficient, physically impossible to manage or exhausting for researchers. AI presents opportunities to make wildlife conservation – and vital components, such as monitoring – more far-reaching and successful. These are some of the projects where AI is currently being used as a wildlife conservation force for good.
Tackling poachers
AI is improving anti-poaching measures in countries such as Zambia. In one project, a 19-km long virtual fence has been created along Lake Itezhi-Tezhi, lined with cameras. AI is being trained to automatically detect boats and to eliminate false readings that can be triggered by birds, waves etc. In protected areas where resources are scarce, this technology can significantly scale up what the rangers here can do.
Detecting and counting animals
One of the biggest challenges for saving animals on the brink of extinction is keeping track of them. Using AI-enabled devices, e.g. drone-based visual AI systems, can support conservation agencies in vast areas – such as the Congo basin, the world’s second largest rainforest. Computer vision technology in the drones can recognise species and provide information about behaviours, as well as monitor whether numbers are expanding or declining. This innovation can be vital in emergency situations. For example, the Queensland University of Technology in Australia developed an AI algorithm that uses drones and infrared imaging to identify koalas by their heat signatures. This was highly effective in finding surviving populations after the 2019 and 2020 bush fires.
Cleaning up marine environments
The waste, especially plastic waste, that we dump in – and close to – our seas presents a huge threat to marine life. However, AI models are now being trained to identify and remove plastic from the water before it can cause a problem. One project uses drones that are trained to identify problematic waste floating in the water, which can then be collected and removed.
Highlighting vulnerability and urgency
Thanks to the use of AI we now know that Brazil has lost more than 15% of its surface water over the past three decades. Factors such as deforestation, economic growth and rising population numbers have all put pressure on water resources but it wasn’t until an AI project analysed water changes across the country that it became clear just how desperate this situation was becoming. The MapBiomas water project processed 150,000+ images generated between 1985 and 2020 by Nasa satellites across the 8.5m sq km of Brazilian territory. It was even able to distinguish between natural and human made bodies of water. Without the AI processing element it simply wouldn’t have been possible to see how urgently action was needed.
