In Australia, deaths from unprovoked shark attacks are the highest in the world. The current rate, 1.5 deaths per year, may seem like a rather insignificant number, but for those living on the island continent who love the water it’s still too many. Until now, many “concerned citizens” have called for some very drastic measures to take care of the problem – from shark nets to baited drum lines. All of these solutions end in only one way: killing sharks for simply trying to survive in their own habitat.
Officials and many conservation advocates have begun to step in and offer up some more humane solutions. By using artificial intelligence, developers have put together a surveillance system that can detect sharks in the water long before they reach the areas in which people are swimming or surfing. While this is a much more humane approach and could keep a high number of sharks from being killed, the root of its inception may actually lie in politics.
Using AI-powered Solutions for Political Gain
For quite some time, Australian politicians and wildlife advocates have gone back and forth over lethal and non-lethal approaches to control the shark population and keep them away from humans in the water. Late last year, a Senate committee report on shark deterrent measures was released to the public in Australia. According to the report, many citizens are beginning to steer away from the usual traps and techniques in favor of more humane approaches. Enter: AI.
The first recommendation of the inquiry is to “immediately replace lethal drum lines” with so-called SMART drum lines and to phase out shark nets. After surveying citizens of Western Australia, the following results were gathered:
- There is little blame on the shark. The tide has turned and the public is sophisticated enough to understand that sharks are not intentionally hurting people.
- There is little blame on the government. Governments that feel they need to continue using shark nets or else face the wrath of the public following a shark bite should rework their political calculations.
- The public no longer supports policies that kill sharks. In WA, 75% supported non-lethal options, in Ballina the number was 83% and in the Sydney experiment it reached 85%.
- A Save the Sharks movement has begun, with the public we have polled consistently voicing greater support for conservation approaches above killing sharks.
- Survey respondents believe that governments choose lethal measures to ease public concern, not to make beaches safer. This is a problem for Australia’s democracy; the public believes that policies are being designed to protect governments, not people.
So now the government in Australia has no choice but to make a change in how it deters sharks from its shores.
Creating Safer Shores
“It is quite obvious that beachgoers and beach recreation will become safer with the shark mitigation technologies, but it is also important for us not to disturb the marine life in general,” said Nabin Sharma, a lecturer at the University of Technology Sydney. “This is a win-win situation for both sharks and humans.”
One of the ways artificial intelligence will help create a better coexistence with sharks is with a fleet of drones. Developed by The Ripper Group, these drones can hover over the shores to detect sharks – or any other type of sea life – in the area. They currently conduct hourly patrols over 40 beaches in New South Wales and eight beaches in Queensland on the country’s east coast.
The drones can fly for up to 28 minutes and spend the rest of each hour on standby. They use an AI algorithm called SharkSpotter, which can tell the difference between objects such as swimmers, surfers, boats, rays, dolphins, and sharks. There is still a lot to learn for these drones, though.
“It is not expected that an AI system will work straight away after deployment, as there are many unknown scenarios,” said Sharma, who worked with Michael Blumenstein and other AI researchers at the University of Technology Sydney. “It will become better and more accurate based on further fine-tuning.”