UNSW and Surf Life Saving Australia have developed RipEye, an AI-powered mobile tool that can help identify rip currents from a photo of a beach, alongside an interactive game designed to help people learn how to spot them.
Rip currents remain one of Australia's deadliest beach hazards, contributing to 47% of the 66 beach drowning deaths recorded in the 2026 National Drowning Report.
Most people know rip currents can be dangerous. The harder part is knowing how to spot one. But what if a photo taken on your phone could help you to learn?
UNSW researchers have developed RipEye, an artificial intelligence (AI) tool that can analyse an image of the surf and identify within seconds where the rip currents may be present.
The technology is part of an Australian Research Council Linkage Project led by UNSW in partnership with Surf Life Saving Australia (SLSA) to help improve rip current awareness and beach safety and potentially save lives.
The tool will be trialled with selected SLSA clubs this summer, with the aim of eventually using it to support public beach safety education and Surf Life Saving patrol operations.
How RipEye works
The tool brings together two complementary tools: AI Live Scan, which identifies potential rip currents in beach images, and 'Spot the Rip', an educational game that helps people develop their own rip-spotting skills.
Dr Mandana Ghanavati, Senior Researcher at UNSW's School of Civil and Environmental Engineering, led the development of the Rip Eye tool, focusing on its performance, accessibility and practicality in the real-world.
An expert in operational AI systems for environmental and public safety applications, she believes the technology could ultimately help reduce drowning deaths.
"Rip currents are dynamic coastal hazards, and beach conditions can change rapidly," Dr Ghanavati says.
"For technology like the RipEye to be useful, accuracy alone isn't enough. The system has to process information efficiently and deliver results while those conditions are still relevant.
"We wanted to make advanced AI simple and accessible. By combining AI-assisted detection with interactive learning, we want to give people the knowledge and tools to make safer decisions before entering the water - without replacing human judgement."
Can you beat AI at spotting a rip?
RipEye goes beyond detecting hazards by using AI to help people develop their own rip-spotting skills.
In an interactive game called Spot the Rip, players progress through different levels, identifying rip currents in images of beaches before comparing their answers with AI. The game provides feedback and offers tips to help players improve their rip-spotting skills.
"The game reflects a broader approach to using AI in public safety. Rather than simply relying on technology to identify the rips, we wanted to use AI to help people develop the skills to recognise one for themselves," Dr Ghanavati says.
The game was designed with some of the groups most at risk of drowning in mind, including young men, international visitors and culturally and linguistically diverse communities. SLSA will trial the game with these groups this summer and gather feedback from beachgoers to help refine the user experience.
"Gamification makes learning more engaging, particularly for younger audiences," she says.
"By offering the game in 11 languages and making it accessible on mobile devices, we hope to break down the barrier to beach safety education no matter where people may be."
Teaching AI to recognise a rip
Rip currents are strong, narrow currents that flow away from the shoreline and can quickly carry swimmers into deeper water. They are considered one of the world's most dangerous coastal hazards, and notoriously difficult to identify.
Scientia Associate Professor Mitchell Harley from UNSW School of Civil and Environmental Engineering is an expert in coastal monitoring technology. Based at the UNSW Water Research Laboratory (WRL) in Manly Vale, he uses innovative technologies to interpret visual data to understand and manage coastal hazards.
Behind the technology is more than 25 years of beach monitoring at WRL, which has accumulated one of the world's largest databases of beach imagery. This project has also benefited from strong engagement with SLSA members, who have contributed rip current imagery, as well as expert and local knowledge to help build a unique dataset for training and testing the AI system.
PhD candidate Shenyang Qian from the School of Computer Science and Engineering used thousands of those images to train the AI model to recognise the visual patterns associated with rip currents.
"I spent the first 12 months of my research mining and annotating through more than 10,000 images of rip currents at different beaches to develop and train the model," says Mr Qian.
"Through this enormous and laborious task, RipEye has learnt to detect rip currents and to adapt its knowledge to changing conditions."
While the tool can correctly detect a rip around 80% of the time, A/Prof. Harley says no detection system will be correct all the time - making it important that users understand that the technology does not guarantee that an area of water is safe.
Australia's diverse coastline has also given researchers a broad range of beach environments and conditions to use in developing the technology.
"Australia is a massive island and depending on where you are, the coastline can be very different," A/Prof. Harley says.
"Rips do not always look the same, so to be able to recognise them across different beaches and conditions can be challenging.
"While the model has primarily been developed using data from Australian beaches, we've also been testing it on images of beaches in Vietnam and California, with promising early results."
The project brings together a broader multidisciplinary team of UNSW academics including Scientia Professor Toby Walsh and Associate Professor Yang Song from UNSW School of Computer Science and Engineering, Associate Professor Amy Peden from the UNSW School of Population Health, and Professor Rob Brander (also known as Dr Rip) and Dr Jaz Lawes (also the Research Manager at SLSA) from the UNSW School of Biological, Earth and Environmental Sciences (BEES).
Putting the technology to the test
Successfully identifying the rip current is only one part of the challenge. The project team also need to understand how the technology can be used safely in the real world.
The tool is currently being tested by SLSA members at specific locations to evaluate how the technology could complement or enhance existing operational procedures. For example, the information from the tool could support decisions about where lifeguards and surf lifesavers place the red and yellow flags marking designated safer swimming and recreating zones on surf beaches.
SLSA plans to implement RipEye to assist in training members and inform patrol operations. Eventually, the goal is to release the tool to the public.
A key question is whether having access to AI assistance could make people overly confident or too reliant on the technology.
Dr Ghanavati says understanding how people interact with the technology is just as important as evaluating its technical performance.
"How people interpret, trust and interact with AI is equally important, yet often overlooked in early warning systems," she says.
"A key part of our research is understanding how AI assistance influences people's decisions and behaviour, including whether it improves hazard awareness or creates a false sense of confidence." 5
An important tool to help reduce rip-related drowning
Australia's coastline has 11,000 mainland beaches, so having active patrol across all beaches throughout the year isn't feasible.
With a particular focus on unpatrolled beaches, the project aims to support the Australian Water Safety Strategy 2030 goal of reducing drowning deaths by 50% by 2030.
Surf Life Saving Australia, National Research Manager and Adjunct Senior Lecturer at UNSW, Dr Jaz Lawes says:
"You can't avoid a hazard if you can't see it. Rip currents are the number one coastal hazard globally and are responsible for more than one in three drowning deaths on Australian beaches, but they can be extremely difficult to identify.
"Our research shows that there is a real gap between confidence and ability: 45% of Australians think they can spot a rip, but only 16% can correctly identify one.
"That's why RipEye is so important. It can help Surf Life Saving turn something that is often difficult to see into tangible information people can understand and use to educate, inform and make safer choices for themselves and their loved ones."
For A/Prof. Harley, the potential impact is what makes the research particularly exciting.
"If AI can help people and professionals better identify the danger in hazards right in front of them - while encouraging them to learn what to look for themselves - it could become another tool in helping beachgoers make safer choices before entering the water."
