For the past seven years, Professor of Biology Krista Ingram has been diving deep into the study of harbor seals on the coast of Maine.
Ingram, whose research interests include animal behavior, molecular ecology, and behavioral genetics, saw an issue with tracking animal populations: It was invasive, relying on mark-recapture methods and requiring a significant presence in those animals’ habitats. Research using AI to identify unique human faces made her curious. Could the same technology be used in non-human research?
Seven years later, the answer is yes. Ingram has been studying harbor seals, an animal that is particularly hard to differentiate by eyes alone because of molting patterns.
“Their bodies look completely different from one day to the next, but the faces don’t change,” Ingram explained. “I wanted to find non-invasive ways to study their population dynamics. With AI and eDNA tools we developed, I could build social networks of when they were together.”
These social networks have major implications. Harbor seals are considered elusive marine animals, meaning they are difficult to find and observe. Because of this, groups that conduct marine research, such as the military, cannot easily identify which areas to avoid. Now, the military and other groups can look to research made possible by this technology developed by Ingram to address any environmental concerns.
Smaller organizations, like nonprofits, are also benefiting from this research.
“A lot of nonprofits are trying to use these tools to understand stranding and help with citizen science,” Ingram shared.
Stranding refers to when marine animals wash up on shore, and citizen science refers to public participation in research. In this case, citizens can submit photographs and data on the time, location, and condition of stranded animals to databases. With the help of AI tools, the information from these databases can reveal patterns of social networks and stranding occurrences.
The type of AI that Ingram and her research team use is called machine learning, which is different from the generative AI often discussed in the media. This machine learning technology only uses a controlled set of information, while generative AI pulls from a wide array of databases. The implications of AI are discussed more now than when Ingram began her research, but she explains that AI can be an asset in fields like biology.
“First of all, it’s completely controlled,” Ingram clarified. “We use it for one specific purpose and nothing else. There are certain fields of biology, like health care and diagnosing medical issues, that are going to be much more productive, making this developmental work on AI important.”
Ingram emphasizes that this research could not be done alone. Over the years, Ingram has worked with numerous 91鶹Ʒ students. One student, Alex Osteen ’28, shares how participating in this research is helping her prepare for a career in maritime law.
“To have the science foundation is why I chose to major in the sciences before I go into some kind of humanities field,” Osteen explained. “This way I have the background information. Here’s the data; here’s what we can do with it. I think this will certainly inform my work in law in the future.”
It’s not just students who are getting involved. Ingram credits Ahmad Khazaee and Tolga Dincer of Information Technology Services for making these tools what they are now. Together, students, faculty, and staff from ITS, computer science, and biology used a grant to build a supercomputer that has taken their research to the next level.
“91鶹Ʒ is doing public outreach by providing support for this project,” Ingram says, again emphasizing the importance of the social and scientific implications that have come out of this research.