Wildlife Cameras Are Accidentally Capturing Humans Behaving Badly

To study wildlife, Dr. Nyeema Harris, an assistant professor in the Ecology and Evolutionary Biology Department at the University of Michigan, uses camera traps — remotely triggered cameras that take pictures when they detect movement and body heat. Harris, a wildlife biologist, is not typically interested in humans, but sometimes they still end up in her photographs.

Between 2016 and 2018, Harris led the first published camera trap survey ever conducted in Burkina Faso and Niger, originally conceived to focus on the critically endangered West African lion. But Harris ended up capturing so much human activity that she expanded the focus of her study to include how humans were using the area. Research on human activity in the wildlife preserve had typically relied on humans reporting their own actions, but with the cameras, Harris could see what they were actually doing. “The data emerged to be a really interesting story that I felt compelled to tell,” Harris says.

Even in studies conducted in remote nature reserves, meant to capture wildlife at its wildest, people showed up.

When camera traps inadvertently capture human activity, it’s called “human bycatch.” And according to a 2018 University of Cambridge study, Harris is far from the only researcher to have ended up with humans in the data. The study included a survey of 235 scientists across 65 countries about their experiences with human bycatch, and 90% of them reported capturing some images of people in their most recent projects. Even in studies conducted in remote nature reserves, meant to capture wildlife at its wildest, people showed up.

As in Harris’s study, this human data doesn’t always stay “bycatch.” Nearly half of respondents to the Cambridge survey said they had used images of people apparently involved in illegal activity to inform wildlife management efforts. Many of them had reported images to law enforcement, others to conservation staff, and some to the media. All this, despite only 8% of projects having set out to capture images of people.

Other camera trap surveys have used the human data explicitly for law enforcement. A 2012 study meant to monitor the endangered Bengal tiger in the Bangladesh Sundarbans captured evidence of poaching. In 2016, authors of the original tiger study conducted another camera trap survey, this time focused explicitly on capturing illegal human activity in the reserve. There are now even camera traps designed specifically to capture images of people.

Ecologists and conservationists use camera traps as a primary tool. Since the traps were first embraced for research in the ’90s, they have helped document a number of rare and endangered species and to establish biodiversity baselines in numerous ecosystems, which are essential for answering questions about the increasingly rapid ecological changes taking place on Earth. Tens of thousands of camera traps are snapping away all over the world. Some researchers envision that an interconnected global network of camera traps will eventually monitor biodiversity around the planet in real time. But the ethics of how to handle the human bycatch that comes with camera trap research are still being debated.

The humans who end up as bycatch aren’t always pleased to have the cameras around. In the Cambridge survey, 76% of respondents reported that people had interfered with their cameras. Cameras were covered or blocked, or had their SD cards removed. Some were stolen, smashed, or shot at. In one instance, a tree was burnt down to eliminate the camera strapped to its trunk.

The most common explanation put forward by researchers for such destruction is that people don’t want to be photographed doing something illegal, or something they consider to be private. Other explanations included theft for resale, outright vandalism, and objections to cameras in “wild” spaces. Researchers suspected fear was a common thread: fear of government, science, surveillance, or some combination of the three.

To protect their equipment, some researchers pay surrounding communities to leave their camera traps alone. Other researchers partner with locals to deploy the camera traps, and share images from surveys with the community to promote goodwill. Many researchers reassure people near the survey area that the cameras are only meant to capture wildlife.

But residents may still have good reason to be wary of how captured images are used.

To efficiently classify the thousands of images accumulated during a given survey, researchers often upload them to a public, online database where volunteers click through images and identify species. That may mean that 20 or 30 strangers see the images.

Dr. Meredith Palmer, a wildlife biologist who manages camera trapping surveys in the United States and in Africa, explains that scientists do take steps to address privacy concerns. She says if an image in one of her projects is classified as having a person in it, it is retired immediately. Some camera trapping databases, like the Smithsonian’s eMammal project, have policies not to include images of humans. Others anonymize the human data by blurring or covering faces.

Palmer is also working on machine learning solutions to the issues caused by human bycatch. “We’re hoping especially to use machine learning to flag images containing people so that those don’t go online at all,” she said. Of course, identifying people with algorithms can work quite the other way. Take this from the Global Wildlife Conservation website: “GWC is working with partners on a range of technology-based solutions including algorithms that can detect individuals of specific species to support intensive monitoring of highly threatened populations, and human identification that enables pictures to be sent in real time to rapid response teams.”

Some scientists believe that blanket rejection of using human bycatch as data would be a missed opportunity. “There is an increasing amount of human pressure and human activities and shared landscapes,” Harris says. “If you’re collecting all this data, even if it’s unintentional, and you don’t use it, and you don’t incorporate it into questions of interest... I think that these kinds of studies give us more information about thinking about strategies that lead to coexistence.”

“I think that as these projects scale-up those conversations will be more important of how we’re using the human data. What does privacy look like? How does that translate into law enforcement? Is it fair?”

Harris hopes her study, for example, might demonstrate how humans and wildlife can coexist in a protected ecosystem, and how the situation is more complex than the typical Poachers vs. Wildlife narrative.

But the human images in Harris’s study were also used by managers of the reserve and by law enforcement, some of whom were co-authors of the study, to inform conservation initiatives. Intentional or not, the science became surveillance.

“I think that as these projects scale-up those conversations will be more important of how we’re using the human data,” said Harris. “What does privacy look like? How does that translate into law enforcement? Is it fair? Is it equitable? Is it justifiable?”

“I suspect that boundary will keep getting pushed back,” says Dr. Chris Sandbrook, a senior lecturer in geography at the University of Cambridge and the lead author of the survey of human bycatch. “It will get pushed back in terms of what is technically feasible. The question is whether society will push back the other way.”