We have a long history of trying to use animals as spies, weapons and warning systems, but the latest plans to use marine organisms as motion sensors may be the strangest yet.
When a beluga whale was spotted wearing a harness recently, some speculated that it had been trained to spy for the Russian army.
That's not as far-fetched as it sounds. Ever since the 1960s, the US Navy has been training dolphins to detect mines and help rescue lost naval swimmers. Russia's been known to do the same.
And sharks, rats and pigeons have been enlisted over the years as eavesdropping devices, with mixed results.
The latest project from the US Defense Advanced Research Projects Agency (Darpa) aims to improve military intelligence by using a range of aquatic creatures - from large fish to humble single-celled organisms - as underwater warning systems.
"We're trying to understand what these organisms can tell us about the presence and movements of all kinds of underwater vehicles in the ocean," says Dr Lori Adornato, programme manager of the Persistent Aquatic Living Sensors (Pals) project.
Living creatures react in various ways to the presence of vehicles. One of the most familiar is the phenomenon of bioluminescence - some marine organisms glow with light when disturbed. This is the focus of one of Darpa's strands of research.
"If you have an organism like noctiluca present on the surface of the ocean and an underwater vehicle that's close to the surface, you will be able to see that from the air because of the bioluminescent trail," explains Dr Adornato.
But the Darpa team is hoping to gain a far more detailed picture of the movements of submarines and underwater drones.
"We want to understand if it is possible to distinguish the response of the organisms to natural versus manmade disturbances, or perhaps even certain types of manmade objects," says Vern Boyle, vice president of advanced programs, emerging capabilities at project participant Northrop Grumman.
"We'll be using advanced processing techniques, including machine learning, to analyse the signals and identify distinguishing features."
The teams are looking at a range of creatures and behaviours. Goliath groupers, for example - which can grow up to 2.5m in length - are known to make a booming sound when approached by divers and also show curiosity when a new object enters their habitat.
"Our non-invasive undersea surveillance and monitoring technologies will be subtly integrated into goliath grouper habitats," says principal investigator Laurent Chérubin of Florida Atlantic University.
"An acoustic response will alert authorities to the presence of a potential threat or intruder, or indeed to any object that is suspicious or out of place within this species' usual visual and acoustic landscape."
These elements of the project involve monitoring what's known as the soundscape, explains Alison Laferriere of project partner Raytheon BBN Technologies. Many species of fish constantly make sound to communicate or in response to external threats.
"If a vehicle comes in to their environment, the thought is that they might change their behaviour in some way that we might be able to detect," she says.
"We're still in the early stages of the project. We recently got back from a trip to the US Virgin Islands where we took some measurements of the soundscape in the presence of a vehicle and without a vehicle present, and we're just beginning to analyse that data now."
Behaviour is an important indicator that potential sub-sea interlopers may be around.
Sea bass, for example, have been observed diving to the bottom of the sea when they hear a loud noise. Might they do the same, in a predictable manner, when encountering an underwater vehicle?
"We have a fairly good feeling that we will see that response, we just need to quantify it," says Dr Helen Bailey, research associate professor at the University of Maryland Center for Environmental Science.
"We can implant miniature depth sensor tags on the fish so we can detect the movement, and there is already the technology in place for that to be a real-time system."
She says there's no reason why an army of black sea bass couldn't provide a cost-effective warning system against enemy subs.
"You've got to compare with the system right now, and the amount of money they're spending on planes, ships, hydrophone equipment, monitoring equipment. All of that gets them very small snapshots, whereas the system we're talking about would last months," she says.
And there's another, even weirder way in which marine organisms can be used to detect the presence of underwater vehicles, says Ms Laferriere.
Snapping shrimp, found all over the world in shallow water at latitudes less than about 40 degrees, continuously snap their claws together, creating a constant sound signal that bounces back off surrounding objects.
As with conventional sonar systems, measuring the time it takes for the sound signal to return, and its strength, can reveal the size, shape and distance of underwater objects.
"The concept doesn't rely on the shrimp changing its behaviour in any way when the vehicle approaches, it just uses the sound it creates," says Ms Laferriere.
This is important because you don't want your surveillance system to be detectable or to make its own noise that interferes with the sensors.
"It's a passive system," she adds. "It will be low-power and capable of detecting even the quietest vehicles."
Why bother using a lot of energy to detect underwater vehicles when you could get a colony of shrimp to do it for you?
But are these projects really feasible or merely the fevered imaginings of grant-seeking researchers?
"There is a global push to work with animals for remote sensing," says Dr Thomas Cameron, lecturer at Essex University's biological sciences school, "both in the case of free living animals or in farming and aquaculture.
"Harnessing the behaviour of animals to give us signals about the environment around us is not new to humans - we have done this with canaries in the mine and with our domestic dogs.
"What is unique in this program is the focus on wild free living organisms and the push to see what we can learn about signalling in the marine environment by focusing on vocal, visual and movement behaviours."
The animal and plant kingdom could be as much a part of "the big data revolution as humans", he concludes.