How to save the coral reef
The rapid decline of coral reefs is not only a critical threat to marine life but also a significant human issue. We spoke to Intel and Accenture to learn how they’re applying AI to help restore the reefs in the Philippines — and what that means for our oceans
Coral reefs are often referred to as the tropical rainforests of the sea. It’s not hard to see why. They’re both naturally beautiful; somehow otherworldly. They’re both home to an incredibly rich diversity of organisms, each an ecosystem that is integral to the life of this planet. They share something else in common, however; something darker: a very real threat to their existence.
While the destruction of the rainforests often makes front-page news — and rightly so — there is less coverage of the shrinking coral reefs in our oceans. And it’s been critical for well over a decade, now.
They may only account for 0.1% of the ocean, but coral reefs are some of the most bio-diverse ecosystems on Earth. A quarter of all marine species are supported by reefs, which can be found on the coastlines of over 100 countries — more than 80 of which are classified as developing.
They aren’t only lifelines for marine life, though. Human life is also at risk if the reefs are allowed to disappear. An Ecosystems and Biodiversity study by the United Nations Environment Programme stated that coral reefs generate up to a staggering $10 billion every year from the likes of fisheries, tourism, coastal protection and medical use. The study also found that 275 million people depend directly on reefs for their livelihood and sustenance, while they are also vital natural barriers against tropical storms — they protect more than 200 million people by eroding 97% of a wave’s strength. Replacing that natural defence by building sea walls would cost $2.5 million per mile.
Reefs are also integral to the production of breakthrough, life-saving medicine: some studies suggest that we are up to 400 times more likely to find new drugs through coral reef ecosystems than through land-based ecosystems.
Yet, while reefs are integral to the planet’s ecosystem, they are degrading at an alarming rate: between 25% and 50% of the world’s live coral has been lost in only the last 30 years. It is a man-made problem: overfishing, unsustainable coastal development, destructive fishing, and climate change causing warming temperatures all contribute significantly to coral reef erosion. And, as ever with global crises, developing countries are particularly at risk.
Some countries decided to do something about it. In 2012, on Pangatalan Island in the western Philippines, in the Pacific Ocean, a group of like-minded locals founded the Sulubaaï Environmental Foundation. A nonprofit, it dedicates itself to conserving, protecting and restoring the natural resources of the local area through sustainable practices and the restoration of ecosystems. In 2016, it created a 46-hectare marine-protected area around Pangatalan. One of the key objectives was to restore the coral reefs that had been destroyed by dynamite fishing — the practice of using explosives to stun or kill schools of fish.
It built the Sulu-Reef Prosthesis, an underwater platform made from concrete to provide support for unstable coral — which then provides a habitat for marine life. It tracked data by diving among the reefs and analysing the photos, but it simply couldn’t take enough images — and with enough accuracy — to build a long-term impression of the reef’s health.
This is where Accenture comes in.
Meet Ewen Plougastel, Managing Director at Accenture Applied Intelligence. Based in Singapore, he leads up a team dedicated to applying AI across hundreds of clients. Outside of work, though, he has a passion shared by several of his colleagues.
“I’m part of the local diving community here in Singapore,” he says. “With a few friends of mine from Accenture we are active members of Accenture Diving Club of Singapore, so diving is something we love — we’ve been diving in the Philippines, Thailand, the Maldives. We love being under the sea, looking at fish, swimming near corals, and enjoying all of the marine life.
“I’m running the AI team, so the idea was to combine our passion for diving, for the maritime environment, with our jobs, which are all about the latest technologies and artificial intelligence. We often have discussions around ‘why are we inventing all of these new technologies, with these new capabilities, if we don’t use them to make the world a better place?’. This should be the first purpose of new technology and inventions: to make the entire planet a better place for all of us, including the oceans. So we wanted to apply all of these great new technologies to something useful and important.”
Plougastel and his colleagues combined their passion and their work when they approached Sulubaaï with a solution to their problem. They helped turn the manual work of diving and photographing — which itself can interfere with the reef’s environment, and divers only have very limited time in the water — and instead placed intelligent underwater video cameras to make the process automated. These cameras are equipped with VASP — Accenture’s Video Analytics Services Platform — which detects and photographs fish as they swim past. VASP uses AI to count and classify the marine life, with the data sent to a surface dashboard, where it provides analytics and trends to researchers in real time, enabling them to make data-driven decisions to protect the coral reef.
“There are two parts to the tech,” says Plougastel. “The first part is getting the cameras under the water, which honestly is quite simple. The second part is applying deep learning and machine learning, to be able to find the patterns and correlations in the images. You combine the weather data, the time data, the tide data, with the activity on the images. When you do this correlation then you find interesting patterns. With AI you can scale that up.”
This is where Intel comes in.
Already regular partners, Accenture and Intel again joined forces in the Philippines and, alongside Sulubaaï, they launched Project: CORaiL.
Patrick Dorsey, Vice President of Product Marketing, Programmable Solutions Group, at Intel, explains their involvement.
“At Intel we have a big focus on social responsibility and making a difference and having an impact on the communities we work with,” Dorsey says. “We looked at the problem CORaiL was trying to solve, and we discovered if we took multiple technologies from the Intel platform we could effectively solve their problems. It was our ability to bring our VPU — our video processing unit — on the front end and our FPGA products in terms of some of the AI analytics, and the platform we have was a very good fit for the problems Accenture were trying to solve.
“You get a massive amount of data on where the fish are going, what type of fish you have, what the flows are, whether they’re nesting there or not. The amount of data, the amount of images that can be collected is not limited by the physical proximity of the divers being in the water.”
Project: CORaiL has so far collected around 40,000 images since it started in May 2019, and that has enabled researchers to gauge the health of the coral reef in real time. This data is crucial to learning how best to restore it.
The project is very much ongoing, and Plougastel’s team is looking to refine its prototype in order to monitor more reefs, and is in talks with several organisations, including an NGO that wants to replicate Project: CORaiL for Australia’s Great Barrier Reef.
Plougastel and Dorsey both believe that developing countries should be the initial focus for implementing AI for social good. But they also know how important it is that they supplement the existing processes they find in each region, rather than riding into each new country like knights in shining armour.
“When we talk to anyone anywhere we learn by listening first,” says Dorsey. “Their way of doing it might be smaller in scale but it can teach us something we can use on a larger scale, and we’re always receptive to doing things differently.
“With Sulubaaï, it was absolutely its expertise. We just came in with the tech; the bricks and mortar to enable what it wanted. It was built all around what data its needs, where the divers need to be, how much of the reef it needed to measure, how expansive the area is, how we make it remote so it can be transferred reliably. It’s a platform that was customised around the problem.”
Plougastel agrees: “We are very keen to help countries that perhaps have more difficulty getting access to this kind of technology. As part of the project we want to educate as well, so we are planning on going into some schools in the Philippines to talk about the project to explain what we are doing, and why. For us, education is very important to what we do.”
Both men believe we’re only at the very beginning of AI’s impact on human life, but with that comes responsibility. Both acknowledge that there remains a pervading fear around its application, and Accenture has a dedicated Responsible AI team that works with clients on how AI can be used responsibly, and how it can be applied for the greater good.
“There’s always a fear that AI will do more evil than good, and my personal view is it’s good to have that fear because then you can watch for it,” says Dorsey. “On a policy level we’re very clear and principled on where AI can be used and how it should be used for good. I think COVID-19 has opened some eyes to how it can be used to be good, but you do have to be careful to do the right thing around the personal freedoms and liberties of people, and we take that very seriously.”
Using tech for social good can only accelerate its societal acceptance, and away from coral reefs both Intel and Accenture have been working on using AI in other socially responsible projects. Intel has implemented AI in Kenya in the fight against illegal poaching, alerting park rangers when poachers are detected in real time.
It has also partnered with relief organisations around the world, including American Red Cross, using AI to improve disaster responses. It built a deep learning algorithm to accelerate the speeds of mapping using satellite imagery.
“We see AI as a big part of solving global issues,” says Dorsey. “Connectivity is also a big part of it, especially with 5G. How do we connect and get information flowing as close to real time as we can. There’s a lot of network transforming happening and how we communicate overall from end points to the core network.”
Accenture is working with farmers in Indonesia to use AI to analyse crop photos taken by drones. This allows the farmer to gauge in real time which crops need more water, or need harvesting, and reduces wastage. Plougastel only sees these kinds of projects and partnerships, including CORaiL, becoming more common as AI is increasingly used for social good.
“We are at the beginning of it,” he says. “There is a huge potential to increase the use of AI in the coming years, and that’s why it’s important now to show that AI can be applied in the right way, and when it is applied correctly it can help make the planet a better place.”
