Students and researchers are increasingly using artificial intelligence to study seal populations more quickly and accurately than ever before. In Scotland’s Ythan Estuary at Newburgh beach, where grey seals gather in large numbers, AI tools are helping turn drone footage into usable data in a fraction of the time required by traditional methods.
Marine biology student Claire Stainfield has spent years collecting drone video of seals on the beach and in the water as part of ongoing research into how the animals use the site. Because seals are sensitive to disturbance, drones allow data collection from a distance without disrupting their behaviour.
Over two years, Stainfield and her team built up a vast library of aerial images showing seals throughout different seasons. In summer, many seals spend time foraging at sea, while in winter they gather ashore for breeding and moulting. Understanding these patterns is important for conservation work and for tracking how environmental change might influence seal behaviour.
But going through this footage by hand identifying seals frame by frame and counting individuals was painstaking. Manual analysis restricted how much data students could process and limited the frequency of surveys.
To speed up the work, Stainfield trained an AI model to recognise seals in the images. She labelled hundreds of photos so the system could learn what a seal looks like against sand, water and rocks. Once trained, the AI was able to quickly scan new footage and identify animals automatically.
The results were dramatic. Where counting seals in a single survey used to take hours, the AI can process the same footage in seconds. This allows students to analyse more surveys more often, giving researchers a richer picture of how populations fluctuate over time.
The time saved also lets students focus on other parts of their research. Rather than spending long hours on manual tagging, they can now interpret patterns, test hypotheses and explore connections between environmental conditions and animal behaviour.
Experts say this approach is part of a broader trend in ecology. Machine learning and computer vision techniques that teach computers to recognise objects in images are being applied to studies of birds, whales, elephants and other species. These tools help scientists make sense of large visual datasets that would otherwise be overwhelming.
Drones themselves have also changed ecological research. They let teams capture high-resolution images without approaching animals directly, reducing stress on wildlife and improving safety for researchers. Combined with AI, drones turn raw footage into a powerful source of data.
Stainfield says one of the most valuable lessons from the project has been learning how to work across disciplines. Understanding how to train, test and refine an AI model has become a useful skill alongside traditional field biology techniques.
She hopes similar AI methods will be used at other sites and for other species. Remote islands, wetlands and rugged coastlines could all benefit from automated image analysis, especially in places where weather or terrain make manual surveys difficult.
Despite the advantages, the technology doesn’t replace the need for human expertise. Researchers must still train the model and check its accuracy, especially in tricky lighting or when animals are partially hidden. Initial setup takes time, but the long-term efficiency gains are clear.
Applying AI at scale could help conservationists track changes in animal populations more responsively. Faster data processing means trends can be spotted sooner and management responses can be informed by larger datasets rather than limited snapshots.
The work at Newburgh beach shows how blending AI with traditional ecology can open new research possibilities. Students like Stainfield are gaining both scientific and technical skills, while wildlife studies benefit from tools that make large datasets manageable.
For seals and other species, smarter data analysis may lead to deeper insights into behaviour, population health and the effects of changing environments. As AI becomes more accessible, its role in ecology is likely to grow, helping future generations of researchers explore the natural world with greater precision and speed.
