The crucial role AI is playing in tackling the poor state of the world's biodiversity has been highlighted at an Abu Dhabi event.
Experts who spoke to The National at the International Union for Conservation of Nature World Conservation Congress offered real-world examples of how AI had helped, but also warned it was not a silver bullet.
Skylight – an AI-driven monitoring and analysis software platform – for example, works by analysing millions of automatic identification systems generated by ships daily and hundreds of thousands of satellite images each week. The AI then combines all the data to help maritime agencies identify events and patterns of interest in often vast and remote seas.
This is all the more pressing given that illegal, unreported and unregulated fishing causing up to an estimated $23.5 billion economic losses annually, according to the UN.
“What AI is doing is filtering through all of that big data,” said Gregg Casad, monitoring, control and surveillance adviser for Skylight. “And finding those incidents and saying, 'here's something of interest that you might want to look at,'” he said.
“A human could look at it, but there's just so much. Where do you start? The sheer fact is we're never going to have enough ships, we're never going to have enough planes and we're never going to have enough boats to go out and patrol all of that.”
Another company harnessing the power of AI is the UAE climate tech venture Nabat, which works with the Environment Agency – Abu Dhabi to restore thousands of hectares of mangroves in the UAE. Nabat uses AI to assess satellite images and other data before drones are sent to seed the ground.
“We are not trying to compete with natural regeneration,” said Taha Ghaznavi, chief product officer at Nabat. “We are trying to extend the boundaries of ecosystems – to reverse the damage that has been done.”
Mr Ghaznavi said about 80 per cent of site restoration projects fail for reasons including poor site selection, cost and time overruns because of the manpower needed, and even injuries. But AI and drone delivery of seeds can help, he said, as they pinpoint specific areas and people are not wading through mud and disrupting ecosystems. Seed survival is put at between 30 and 50 per cent after six months, but he said it is complex.
“Our AI is ecology-trained,” he said, calling it “tech with muddy boots” – he says the company has more ecologists than technologists. It is very much a human-machine partnership and it should remain this way.”
The event in Abu Dhabi showed the pressure the natural world is under. The latest IUCN Red List of Threatened Species, which was issued on Friday, showed that more than a quarter of the animals, plants and fungi assessed are at risk. The reasons for this are manifold, and include climate change and habitat loss.
Wildlife trafficking is part of the problem, with a 2024 report from the UN Office on Drugs and Crime stating that more than 4,000 species are affected. Another AI-driven solution to try to tackle this was also showcased in Abu Dhabi. Earthranger was developed in partnership with Save the Elephants and other conservation groups.
Global populations tracked include 25 per cent of elephants, 10 per cent of rhinos and 50 per cent of scimitar-horned oryx, according to a presentation given at the congress. This enables the monitoring of unusual events around camera traps, such as unwanted human activity or the presence of endangered animals, in a much more efficient way than is possible by sifting through scores of photos individually. It also zooms out to identify broader patterns and could help predict future events.
“Imagine a camera hidden in a tree,” said Jes Lefcourt – director at Earthranger. “We are monitoring 24/7 and can also see long distances and see in the dark. And if a human walks by in one of these very remote places that people generally shouldn't be, then it will send a signal up to the cloud.”
This alerts the dashboard in an operations room, allowing officers to send rangers to the site immediately. Hundreds of sites across the world from South Africa to Kenya have camera traps that are integrated into Earthranger looking for poachers. Mr Lefcourt didn't want to give out sensitive details about locations, but said poaching could be “reduced significantly” using this.
AI, therefore, is analysing data and alerting people faster than could be done before but is it overhyped? “There are many examples where you see AI that's been added to a press release and question whether or not it is even using AI, but I think there's also an enormous amount of potential that we haven't even realised,” said Mr Lefcourt.
Both Skylight and Earthranger are part of the AI2 umbrella – a Seattle based non-profit AI research institute founded in 2014 by the late Paul Allen, co-founder of Microsoft. When asked if AI was truly making a difference, Mr Casad said: “It absolutely is. These tools are providing people in the field with information in a time frame that they can act on.”
Others were also showcasing the power of AI at the IUCN. Eric Schmidt, executive director of the US non-profit Wildlife Protection Solutions, outlined how its technology can also defend against poaching.
Its AI-driven software scans more than 75,000 images a day from cameras and CCTV among others in protected areas to identify threats and can then send alerts to rangers or wardens through SMS or WhatsApp, reducing reliance on time-consuming direct observation.
“On an almost a daily basis [it is] catching intruders coming into places that they shouldn't be,” said Mr Schmidt. “[I’ve had] people in the US monitoring cameras on the app and they see something that looks suspicious, hit a button in the app that automatically alerted the local warden in Africa who said, ‘yes, indeed, that is a poacher’, and then he activates his response team,” he said.
“So it's kind of a cool scenario where the technology also helps unite people around the world in a conservation effort and a positive outcome.” Mr Schmidt cautioned, however, that AI is just a component of the efforts to tackle possible shady activity.
“You need really strong people processes to take the data that it is presented to you and act on it,” he said. “If you just put it out there and expect it to make a difference, it won't.”
There is also another environmental side to AI. Data centres that house servers can lead to electronic waste, they consume water and use increasingly large amounts of electricity often generated by fossil fuels, which create warming greenhouse gases.
This is a point noted by Joe Walston, executive vice president for global programmes at the Wildlife Conservation Society, a global conservation organisation active in more than 50 countries.
“We will have extraordinary achievements, through AI, in food production systems and clean technologies, on reductions in pollution, and on the crucial decoupling of economic growth from environmental degradation,” he said. Mr Walston said it would also produce tools that create “novel forms of pressures on the environment” and make “vast demands on energy production”.
“It will come down to human ingenuity, collaboration and human goodness to ensure that the balance is in nature’s favour.”
