9 June 2022
Innovation in almost any sector is advancing at a pace that we have not seen before. This new blogpost series explores these innovative sectors in pursuit of creating a liberal environment, to push the boundaries of science even further.
By Elias Rosell, Political Editor for Östersunds-Posten.
None of the 20 biodiversity targets for 2011-2020 was fully achieved by 2020, which is a problem for more than just bird watchers and hikers.
Birds, bees, and other pollinators account for about a third of the world’s crops. Seafood is the most important source of protein for three billion people. However, a third of the global stock is overfished. About 60-80 per cent of antibiotics and anti-cancer medications originate from natural substances. Despite the advantages of modern chemistry and pharmaceutical applications, most modern drugs take advantage of the properties found in such compounds. If more species are lost, we risk missing out on the development of future medicines.
These are just a few reasons why we have to do better in conserving biodiversity and healthy ecosystems. One way to improve the task of preserving natural intelligence is to make use of artificial intelligence. Around the world, there are promising examples of how AI monitors biodiversity, protects animals, and identifies areas worthy of protection. Let us take a closer look at AI’s applications in this regard.
When it comes to biodiversity, camera traps are perhaps the most obvious tool for keeping track of different species both on land and under the sea. They consist of cameras placed in selected observation points in nature which, with the help of infrared sensors, detect when an animal or some other warm object is approaching and then take a picture. They provide scientists with millions of pictures to analyse. However, doing so can be very time-consuming, to put it mildly. Thankfully, artificial intelligence can be used to filter out blank images and classify species.
One example of such a project is Wildlife Insights, involving Google and other companies. Anyone can use Wildlife Insight to upload photos from a camera trap for analysis. Wildlife Insights can analyse 3.6 million photos per hour and has so far learned to identify almost a thousand different species.
Artificial intelligence can also be used to analyse audio recordings. “By placing a few microphones in a forest, you can get exposure to a lot of different organisms occupying that forest without needing to detect them during fieldwork,” says Rajesh Sankaran, an experimental systems specialist at Argonne National Laboratory. Before using sound recordings and AI in the wild, they will allow the algorithms to train themselves at a zoo. The hope is to use the technology to identify which species are present in a given area and which are missing.
With better knowledge about biodiversity, policymakers and stakeholders can make better decisions on how best to protect it.
In addition to species identification, AI can also be used to protect animals from various threats. A potential danger to flying animals is wind turbines. Even if opponents of wind power sometimes exaggerate the risks – for example, domestic cats cause significantly more dead birds than wind power plants do – wind farms placed in the wrong place can threaten, for example, birds of prey. Through a system that combines cameras and AI, endangered bird species can be detected when they are one kilometre away from a wind turbine, which means that the turbines can be slowed in speed or even shut down. Thanks to this, it is estimated that golden eagle fatalities could be reduced by 82 per cent.
But conserving biodiversity is not just about protecting nature. Sometimes, it can actually be about eradicating animals that have ended up in the wrong place. Invasive alien species pose a major threat to biodiversity. Since the 17th century, they have “contributed to nearly 40 per cent of all animal extinctions for which the cause of death is known,” according to one article in the journal Sustainability.
One such species is the mosquitofish, which has caused great damage to ecosystems around the world by attacking local fish species and tadpoles, among other things. At the same time, the mosquitofish has proven to be difficult to get rid of once it has established itself somewhere. Biologists and engineers have developed a robotic fish that can identify mosquitofish with the help of artificial intelligence. The robot attacks the mosquitofish when it approaches a tadpole. Experiments in a test environment show that mosquitofish then become stressed and have reduced fertility when the robotic fish is around. There have been no experiments in the wild as of yet, but this could potentially be a way to aid biodiversity through artificial intelligence in the future.
In addition to protecting specific species at certain times, a new study shows that AI can be used to determine which areas should be protected.
Researchers from Sweden, the UK and Switzerland have let an algorithm play the role of policymaker, as in a video game, where the reward is the number of species spared from extinction at the end of the game.
After a lot of training, the system has become better at determining which areas should be protected to maximise biodiversity according to a given budget than other, simpler conservation policies.
Another, possibly more indirect, way of using AI to help biodiversity is to make agriculture more surface-efficient and environmentally friendly.
On a global level, habitat loss due to the expansion of agriculture is one of the biggest threats to biodiversity. Of the world’s habitable land, about half is used for agriculture.
In Europe, though, the picture looks better. Since the 1960s, the amount of land used for agricultural production has decreased in many European countries – partly because agriculture has become more productive. When less land needs to be used for agricultural purposes, there is more space for wildlife and natural spaces. Decreasing the amount of agricultural land, therefore, is one reason for the remarkable comeback of European mammals that we have seen in recent decades – for example, the population of European bison increased by 3,000 percent from 1960 to 2013, and the grey wolf increased by 300 percent during the same period.
Therefore, more efficient agriculture is a key to increased biodiversity in Europe, not to mention the rest of the world. Artificial intelligence can make agriculture more productive by ensuring that water and pesticides are being used more efficiently and by drones that monitor for pests. More efficient use of pesticides can also reduce the damage that agriculture inflicts on insect biodiversity.
So far, the EU has not been focusing on how AI can be used for biodiversity in a major way. It is difficult to say exactly why, but one possible explanation is that biodiversity has been overshadowed by the issue of climate change.
The main focus in the climate and environment chapter in the European Commission’s “COORDINATED PLAN ON ARTIFICIAL INTELLIGENCE 2021 REVIEW” is on climate and energy; biodiversity is not mentioned.
The “EU Biodiversity Strategy for 2030” does not contain anything related to how digital tools such as AI can be used for biodiversity.
The development around AI towards biodiversity has only just begun, and we do not know exactly which future applications we can expect. The Commission and policymakers at both EU and national levels must continue to follow developments and should be aware of AI’s potential for increasing biodiversity.
In that vein, the EU can contribute by using research funds towards the study of AI and biodiversity applications; beyond that, EU policy should become technology neutral.
Using AI for biodiversity is not a goal in and of itself. Ultimately, it should not be up to the EU or politicians to decide precisely which technologies should be used to protect biodiversity.
What the EU ought to do is to create better conditions for companies and other organizations to innovate when it comes to AI and other technologies. Therefore, the EU’s digital single market must be completed.
The EU and its Member States should also pursue an ambitious policy to safeguard biodiversity. Such a policy could increase the demand and create a market for artificial intelligence in order to protect biodiversity. One example would be stricter requirements on wind turbines so as not to negatively affect avian biodiversity. The response from wind power producers could then be utilizing AI cameras to recognize birds of prey and then slowing down the turbines as needed. Those wind power producers who wish to use a different technology to protect biodiversity should be welcome to do so – as long as the desired result is achieved.
A big stumbling block regarding the EU’s approach towards biodiversity is agriculture. The EU should have technology-neutral instruments to make agriculture more efficient and environmentally friendly. It is unfortunate that the system of direct payments to farmers on a per-hectare basis will continue, albeit with some changes, in the new CAP applying from 2023. If the EU instead were to implement stronger incentives for more efficient agricultural practices, the demand for using artificial intelligence to that end could increase.
As long as protection from privacy breaches can be ensured, companies, NGOs, and research institutions should have a great deal of freedom to develop exciting new uses for AI in the pursuit of biological diversity. The best applications are still ahead of us.
Elias Rosell is an Earth Scientist who is working as a journalist, he often writes about environmental and energy issues. He now works as a political editor at Östersunds-Posten and has previously worked as a research coordinator at the European Liberal Forum.
DISCLAIMER: Published by the European Liberal Forum. The opinions expressed in this publication are those of the author(s) and do not necessarily represent those of the European Liberal Forum.