In 2016, a bright yellow robot began moving up and down the Great Barrier Reef, hunting down a particular type of starfish and moving in for the kill: After the robot injects the animal with poison, it dies within a day.
The starfish feeds on coral, and even though the animal is native to the reef, an out-of-control population has made it a threat to the reef’s survival. In theory, the robot, which is still in development but has been successful in pilot tests, can help solve the problem. It’s one example a growing number of bots designed to reshape ecology.
An upcoming journal article asks how automation might ultimately go further: A “wildness creator” might create and maintain “wild” places without any human intervention.
“We focused on this idea of wildness because it seems like it’s an interesting way of, in a way, poking back at what we think nature is,” Bradley Cantrell, an associate professor of landscape architectural technology at Harvard Graduate School of Design and TED fellow, who worked on the research with ecologist Erle Ellis and environmental historian Laura Martin, tells Co.Exist. “Could there be wildness that’s highly managed through technology?”
Because of the scale of human impacts on ecology–from building cities and razing forests for agriculture to pollution and climate change–ecological restoration and conservation is also becoming increasingly difficult to manage. Automation might help.
The paper looks at several examples of semi-autonomous ecological projects that are already happening or in development. One startup is using drones to replant trees at rates faster than humans could. A “virtual fence” along Australian roads automatically warns wildlife when a car is coming. MIT is developing swarming robots that could clean up oil spills.
Other projects are using technology to monitor wilderness and human impacts; drones can detect early-stage wildfires, track wildlife populations or pollution, or catch poachers. Satellites can track illegal fishing boats and illegal deforestation.
Using that data, analysis software can determine the best locations to intervene to restore an ecosystem.
The paper considers how several technologies might work in a fully autonomous system, using machine learning to choose and implement the best interventions. In a coastal wetland, a system might identify human influences such as pollution, noise, or waste, and then automatically act to respond to those influences to help local plants or animals, learning as it goes which strategies help most.
In time, driven by automation, the “wild” area might look unrecognizable to humans–and the researchers ask if that’s a ethical system to create. Even in smaller interventions, such as COTSBot, the starfish-killing robot, automation might not be the ideal solution.
“We were trying to conserve the Great Barrier Reef, and we’ve come to the conclusion that we have a species that’s eating the reef,” says Cantrell. “What we’re not really looking at with that is that species is there because of these huge nutrient plumes that are coming off the rivers because of our agriculture upstream–we’re kind of ignoring that aspect.”
If deployed in the right way, however–and perhaps not to the degree of the “wildness creator” proposed in the paper–Cantrell thinks that automated systems could be useful. “It’s more about us understanding our relationship and interaction with the environment as opposed to us sequestering the environment away from ourselves–I think my take would be that it requires a much more nuanced view of human and other species interaction.”
Future cities and infrastructure, he says, may be better connected with the natural context in which they’re built, and in that setting, these technologies could play a role. “It’s not about controlling that system, but about, in a way, choreographing the system,” he says.