AI, citizen science combine to help save sharks
The platform uses artificial intelligence to turn shark photos posted online into real-time data that helps scientists monitor and protect threatened shark populations worldwide.
While in Monterey Bay, a basking shark was seen while whale watching. It's photos like this that can contribute to building a better database on sharks. Photo courtesy of Cheryl Butner.
One-third of shark species are at risk of extinction, yet scientists still lack basic data on their habitats, populations, and trends.
To solve this, researchers at Virginia Tech, Stanford University, and others are building the world’s largest open database of shark sightings using online photos called sharkPulse.
Instead of relying on individual submissions, the platform described in the research paper uses artificial intelligence (AI) to scan online sources for shark photos, automatically extracting location data, timestamps, and species identifications.
The research was published in Fish and Fisheries on Aug. 11.
Images are validated by both the public and experts, then added to a searchable database. This approach allows researchers to map shark populations and track changes in abundance and distribution with unprecedented scale and speed.
“This shifts citizen science from voluntary submissions to intelligent autonomous discovery, turning everyday digital activity into conservation data,” said Francesco Ferretti, the study’s lead author and assistant professor in the Department of Fish and Wildlife Conservation.
The current data gap hinders conservation as researchers don’t know the all the habitats and areas that sharks frequent.
“With cameras in nearly everyone’s hands, our encounters with the ocean are being recorded more than ever,” said Jeremy Jenrette, a Ph.D. candidate in the department and author on the project. “SharkPulse taps into this unprecedented global stream of images and videos, using AI and data science to passively monitor shark populations at a scale never before possible.”
Utilizing this data source offers a practical way to address existing knowledge gaps in the study of threatened marine species, contributing to more informed and coordinated conservation efforts.
“We can’t protect what we don’t know,” Ferretti said. “From our findings outlined in the paper, sharkPulse turns scattered signals into knowledge.”
This finding builds on Ferretti’s previous global shark studies.
- Another project launched the “White Shark Chase” to track critically endangered white sharks in the region. Using environmental DNA and baited cameras, researchers detected white sharks at several sites in the Sicilian Channel. The findings support efforts to launch a monitoring program aimed at preventing the population’s extinction.
- During the expedition, researchers tagged a juvenile shortfin mako shark — the first recorded tagging of its kind in the Mediterranean. Tracking showed the shark traveled more than 750 miles in 54 days, underscoring the species’ vast range and the need for broader conservation planning.
- Virginia Tech also contributed to MegaMove, a global effort to track more than 100 large marine mammal species and map vital habitats and migration routes. Findings showed that 60 percent of these critical habitats fall outside protected areas, pointing to the need for wider conservation measures.
Together, these projects reflect Virginia Tech’s dedication to marine conservation through targeted research, technology, and global collaboration.
Tracking global trends
sharkPulse automates data collection, reducing reliance on manual uploads — a shift from traditional citizen science. It still relies on public participation, though, for validating sightings and training AI models.
To date, sharkPulse has validated more than 91,000 records across 285 shark species — nearly 53 percent of known species. It’s helped identify new shark hotspots, such as white sharks in the Mediterranean, and can support The International Union for Conservation of Nature’s Red List of Threatened Species assessments by generating dynamic distribution maps and abundance trends that update as new data are incorporated. The union is a comprehensive information source on the global conservation status of animal, fungi and plant species
Impact across the mid-Atlantic
In Virginia, bull sharks are considered summer visitors in the Chesapeake Bay, yet little is known about their movements or population size. On Aug. 13, 2018, a 2.6-meter bull shark was caught and photographed by a Menhaden fisher off Cedar point in St. Mary’s County.
“sharkPulse is built around this kind of records. The platform gives us a way to collect and organize fugitive local information and transform them into scientific knowledge,” Ferretti said. “It strengthens our understanding of marine ecosystems close to home — not just globally.”
Virginia Tech students and faculty from across disciplines have contributed to the project, including computer science, wildlife conservation, and data science. The team is pursuing new grants to expand the platform’s reach and sustainability.
The waters ahead
The researchers hope to scale sharkPulse further, using multilingual data mining and international partnerships to fill geographic gaps. They’re exploring how to make the data more useful to policymakers, fishery managers, and conservation groups.
“This is about creating an always-on pulse monitor for the ocean,” Ferretti said. “The more we see, the more we can do to protect.”
The project is a cornerstone of Jenrette’s Ph.D., where he has been tasked with building practical, scalable tools for shark conservation.
“I actively develop and refine sharkPulse, which has allowed me to integrate machine learning, big data pipelines, and citizen science into a single framework aimed at addressing a pressing challenge in marine conservation: how to monitor wide-ranging, poorly understood shark populations in near real time,” he said.
It’s also flexible. The team sees the sharkPulse model as a blueprint to adapt the technology to other species groups from sea turtles to bats.
Together with Ferretti and Jenrette, the study also included co-authors Stefano Moro of Stazione Zoologica Anton Dohrn; Cheryl Butner, Fiorenza Micheli, and Trevor Hastie of Stanford University; Edward Fox of Virginia Tech; Steven Haddock of the Monterey Bay Aquarium Research Institute; and Salvador Jorgensen of California State University of Monterey Bay. Financial support was provided by the Bertarelli Foundation, the Virginia Tech Global Change Center, and the Lenfest Ocean Program.
Original study: DOI doi.org/10.1111/faf.70006
