(CNN) — A hidden rip current in the shallows; a speedboat too close to shore; a swimmer out of their depth. These are just some of the dangers beach lifeguards constantly watch out for to keep beach-goers safe.
But drowning is still a major problem; according to the World Health Organization, there were 320,000 drownings worldwide in 2016, making it the third-highest, injury-related cause of death.
Now, an Israeli startup says it has a solution. Sightbit has created an AI lifeguard -- also called Sightbit -- that helps human lifeguards watch the beach and identify potential hazards, says the company's co-founder and CEO, Netanel Eliav.
The camera-based system monitors the environment, assesses risk and detects dangers -- sharing real-time information with lifeguards on duty, so they can act early to prevent fatal accidents, says Eliav.
Sightbit tested its AI lifeguard system at Palmachim Beach near Tel Aviv, Israel.
Sightbit LTD, Israel | Netanel Eliav
A lifeguard's job is to prevent incidents around a pool, on the beach or in the water, and to respond in emergencies. But a lifeguard can't watch everything all the time, and that's how accidents happen, says Eliav. "About 90% of the work of any lifeguard is scanning, and just trying to be one step ahead," he says.
This is where Sightbit comes in. "Computers are usually better than humans at analytics and scanning," says Eliav. He claims the AI camera system identifies water hazards 80% faster than a person, making lifeguards aware of risks before there's an emergency.
The AI lifeguard units consist of three cameras, each covering a strip of beach 100 to 150 meters wide. The system notifies the human lifeguard on duty when it detects a risk, such as an unattended child at the water's edge, or a jet ski in the swimming area. Unlike human lifeguards, the camera is able to see everything in its vision area simultaneously: this saves precious time and could mean the difference between life and death, says Eliav.
The AI lifeguard flags potenital risks to human lifeguards, based on deep learning and risk assessment algorithms.
Sightbit LTD, Israel | Netanel Eliav
Sightbit is especially useful for crowded beaches, he says, allowing larger areas to be monitored without increasing the number of lifeguards.
Some fine-tuning is still needed. Tests carried out on Palmachim Beach near Tel Aviv, Israel, revealed that the system is over-sensitive and issues too many "false positive" alerts, he says, so the company is adjusting the risk assessment algorithms. "The next step is to develop the product to be accurate enough and independent enough to use on unguarded beaches," he says, adding that potentially, multiple unmanned beaches could be monitored by Sightbit, supported by first-aiders stationed in a centralized location.
Smart beaches for smart cities
Sightbit isn't the only tech initiative designed to make beaches safer. Shark Smart, an app created by the government of Western Australia, allows users to report shark activity in surfing and beach hotspots. In New South Wales, Smart Beaches -- a collaboration between the Northern Beaches Council, Lake Macquarie City Council, and University of Technology Sydney -- gathers data to "improve lifeguard services," says project manager Tony Blunden.
The Smart Beaches initiative attaches GPS trackers to rescue paddleboards, helping lifeguards collect data on their daily activities.
Lake Macquarie City Council
Lifeguards typically log activities and incidents at the end of the day, which can lead to inaccuracies because it relies on human memory, says Blunden.
Crowd monitoring cameras take the guesswork out of reporting visitor numbers, while GPS trackers attached to equipment like jet skis and rescue paddleboards enable lifeguards and beach managers to monitor how and when the equipment is used.
This data is tied in with information on weather and sea conditions, revealing environmental risk factors that lifeguards should be aware of, such as the link between low tide and increased risk of drowning.
Blunden hopes the project, which is underway at four beaches, will help lifeguard teams optimize how they operate.
He emphasizes that Australia already has "good, experienced lifeguard services" -- the technology is designed to support lifeguards, not replace them. The aim, he says, is "to help the right people get the right equipment, at the right place at the right time."