At Padaro Beach in California, a new high-tech project called SharkEye is using drones and artificial intelligence to keep beachgoers safe from juvenile great white sharks. With sharks becoming a common sight near the shore, especially in the summer, SharkEye’s drones capture video footage and use AI to spot these predators more accurately than human eyes alone. This technology is helping to send real-time alerts to lifeguards, surf instructors, and parents, aiming to prevent potential shark encounters. The AI model is showing promising results, potentially making beach safety smarter and more efficient as it evolves.

Local children enjoy getting together at California’s Padaro Beach on summer mornings to learn how to surf in mild whitewater waves. A few years back, young great white sharks started frequenting the beach as well.

This resulted in the establishment of SharkEye, a program at the Benioff Ocean Science Laboratory (BOSL) at the University of California, Santa Barbara that uses drones to observe underwater activity.

The eighty or so people who have signed up for alerts from SharkEye, which includes local lifeguards, surf shop owners, and parents of students taking lessons, receive a text message if a shark is seen.

In other recent efforts, lifeguards and officials from Sydney to New York have been utilizing drones to watch live video transmitted from a camera to ensure the safety of beachgoers. That means a pilot must maintain visual focus while navigating turbulent waters and sunglare in order to distinguish sharks from paddleboarders, seals, and rippling kelp strands. According to one study, drones under human supervision only identify sharks roughly 60% of the time.

SharkEye is a program that combines community safety and research, and it uses the video it gathers to study shark behavior. To teach it to spot great white sharks close to Santa Barbara, the company is also feeding its footage into a computer vision machine learning model. This is a sort of artificial intelligence (AI) technology that allows computers to extract information from photos and videos.

“Automating shark detection … can (also) be really helpful for a lot of communities outside of ours here in California,” Neil Nathan, a project scientist with BOSL, who graduated from Stanford University with a master’s degree in environmental studies a few years ago, told CNN.

Human vs. AI in Shark Detection

Drones are becoming more and more popular, and social media is spreading, giving the impression that sharks are everywhere. The fact that young great white sharks, which may reach lengths of eight to ten feet, prefer to congregate close to the coast, increasing their visibility to beachgoers, and that rising ocean temperatures are forcing sharks into new habitats are also contributing factors.

Shark attacks are uncommon, though. Unprovoked bites affected 69 individuals worldwide in 2023, which is comparable to the 63 yearly occurrences that occurred on average between 2018 and 2022. According to the International Shark Attack File maintained by the Florida Museum of Natural History, only ten of them perished.

While no fatal attacks have been reported at Padaro Beach, the presence of sharks in the area alarmed several locals.

SharkEye has been conducting drone flights to monitor the coastline regularly for around five years. In one instance, the company saw fifteen young great white sharks in a single day.

According to Nathan, preliminary testing shows that the AI system is already operating “incredibly well,” identifying the majority of sharks that a human can identify and occasionally sharks that a human missed, possibly because it was swimming too deeply to be easily spotted.

The group started putting its technologies to the test in the field this summer by competing AI against drone pilots. As she scans the region, its pilot tallies the sharks she sees. Next, the SharkEye model examines the footage to determine the number of sharks it can detect.

Currently, human analysis provides the foundation for community notifications. If all goes according to plan, Nathan indicated that by the conclusion of the season or the beginning of the following summer, such reports might be helped by AI in addition to manual monitoring and checks. The procedure might perhaps become fully automated in the future, which would speed up and possibly improve accuracy.

Artificial Intelligence and Wildlife

AI technologies are being used in a variety of ways to reduce conflict between people and wildlife. AI-enabled cameras are warning villagers in India when tigers are approaching their cattle, while technology is being utilized in Australia to control some of its more hazardous inhabitants.

The world’s first shark detection algorithms, according to Ripper Corp and researchers, were developed a few years ago and used in drones. To detect sharks and crocodiles, the most recent version of the software is being tested in the Caribbean, Mexico, and the Australian state of Queensland.

However, shark detection does not yet make extensive use of AI. Drones are used at 50 places by Surf Life Saving New South Wales, which guards numerous beaches throughout the coast of the state, including the well-known Bondi Beach in Sydney. However, a representative informed CNN that AI is not currently being used by their drones.

In 2022, researchers from one Australian institution that developed AI-enhanced shark-spotting technologies noted that the technology can have issues when it comes to situations that weren’t included in the training set.

SharkEye intends to develop an AI-powered program that is simple for users such as lifeguards and drone enthusiasts to run their footage through, in addition to making its model freely available for researchers to modify or expand upon. That might improve human safety while simultaneously advancing our knowledge of and ability to safeguard sharks.

According to Nathan, it’s unclear how much retraining will be necessary before SharkEye opens up shop in other areas. In areas of California with similar coastlines, he hopes drone operators won’t encounter too many problems if they fly at the same speed and altitude.

According to local media, Honolulu officials stated this month that they are thinking of starting a drone shark surveillance program. More retraining might be required if SharkEye’s technology were to be applied in locations like Hawaii, where tiger sharks are the main threat and the water’s color varies. Nathan, nevertheless, stated that SharkEye is willing to collaborate with other communities to assist modify the model.

“Communities want to have that knowledge and that awareness so it’s easier to more safely share the water with these creatures,” said Nathan. “Sharks are an incredible species that we still are always learning new things about.”

Last month, GreatGameInternational reported that training AI models like Google’s Gemini Ultra, costing $191 million, and OpenAI’s GPT-4, costing $78.4 million, is becoming increasingly expensive, as shown in Stanford’s 2024 AI Index Report.