Mercury Logo

Hello, we’re Mercury. We’re the banking stack for startups. We offer bank accounts with all of the tools needed to grow a business — virtual and physical cards, wires, currency exchange, and more.

We appreciate diving into a good read with a cup of tea, so grab your favorite beverage and enjoy Meridian.

Meridian Logo

The researchers building the internet for animals
The researchers building the internet for animals

Tech around the world

The researchers building the internet for animals

A group at the fringes of academia are hoping to make animal-computer interaction take off and change how we interact with our pets, livestock, and environment.

Written by Chris Stokel-Walker

Illustrations by Ryan Nguyen

In 2011, when Ilyena Hirskyj-Douglas was 21, she got a black Labrador named Zack. It was her way of making liminal student life feel more like home while she was studying for her master’s degree in computer science at the University of Central Lancaster.

She spent a lot of time watching television. One day, she noticed something strange. “My dog was watching TV too,” she joked with her housemates. But the longer she watched her dog, the more she wondered if it wasn’t just a joke.

So she built a system that could recognize her dog’s presence and allow it to control what was on the television screen. She then used this data to better understand how the dog interacted with the screen. And before she knew it, she was pulled into the complex and emerging field of animal-computer interaction (ACI).

“I would often ring my dad and we would show our dogs to each other,” says Hirskyj-Douglas, who’s now an assistant professor in ACI at the University of Glasgow, Scotland. Her Ph.D. work focuses on canine interactions with screens. “I began thinking about what a dog internet would be like. With more dogs being left alone and having behavioral issues, I thought the internet had a lot to offer.”

Hirskyj-Douglas has since developed the DogPhone, which allows dogs to move a small ball to coordinate with a video phone. If the dog picks up the ball, the DogPhone video calls its owner. If the owner calls the dog, the dog can pick up the ball if they want to respond. As Hirskyj-Douglas explains in her paper, the roadblock she ran into was having to train a dog to use the system — the next iteration might fit into a dog's daily schedule (or existing toys) more seamlessly.

Monkey Tunnel was born out of Hirskyj-Douglas' attempt to solve a similar problem: she was working with white-faced saki monkeys at university and became intrigued with how monkeys controlled their environments. So she built an interactive computer for primates to toggle between audio and video, and choose what they wanted to tune in to — rain sounds, videos of worms, underwater scenes.

“We developed a system, like for the dogs, that recognizes their behavior and uses it to trigger computers,” she says. “We looked into the monkeys controlling audio and video to turn on and off and pick between different stimuli.”

DogPhone and Monkey Tunnel have both remained academic projects; Hirskyj-Douglas is now back in the lab. ACI is an incredibly hard field to commercialize. One reason is that the field is so early.

It wasn’t until 2011 that Clara Mancini, professor of animal-computer interaction at the Open University, penned one of the first manifestos to define the field. In the piece, Mancini suggests that ACI could revolutionize how we interact with animals, from helping us better fulfill their psychological and physical needs to giving farm animals control over the processes we include them in.

And the first ACI academic conference devoted solely to the area didn’t take place until 2016 — before that, ACI only got a one-day affiliated event at another conference, the International Conference on Advances in Computer Entertainment Technology.

Unlike the breakneck progress in human-computer interaction (HCI) — a field of academia that has been around since the 1980s and has influenced everything from how our phone screens are displayed to user-friendliness in websites — ACI has moved at a snail's pace. HCI has moved so fast that user experience (UX) has developed as its commercial counterpart, a design practice that continues to dip back into what academics are learning. On the other hand, barring an eye-catching trial of Dog TV in 2012, little has happened commercially for ACI.

It’s not that there isn’t a demand for technology that interacts with animals, particularly our pets — Lerer Hippeau pegs the size of the pet tech market at $99B. And in 2021, pet tech companies raised record amounts of venture capital, surpassing $1B in total.

But none of the biggest deals were rooted in ACI and its commercial manifestation, animal tech. Most startups funded focused on physical care — pet insurance companies like ManyPets, which raised a $350M Series D, pet DNA startups like Embark, which raised a $75M Series B, and vet care companies like Bond Vet, which raised a $170M private equity round.

Here and there in the flurry of tech coverage, one might catch a glimmer of ACI. Like at CES in 2022, when FluentPet introduced talking buttons that pets could use to communicate. Or on Kickstarter, which has helped launch a small number of successful animal tech startups, like Ukraine-founded, San Francisco-based Petcube, which develops animal-specific software and cameras that allow pets to interact with different treats and communicate with their voices. After its 2013 Kickstart launch, PetCube went through Y Combinator in 2016. It has since raised a total of $14M in funding.

But most animal tech companies don’t get as far as PetCube.

“It’s hard to find investment,” admits Yaroslav Azhynuk, CEO and co-founder of Petcube. He believes there’ll be a flourish in investment in the future once the technology matures, but it’ll take some time — and ACI will always lag behind what we build for humans.

“By definition, it’s a smaller market,” he says. He believes that it’s the same problem that the broader tech industry often faces. Although we brandish and flash our iPhones with aplomb, we rarely think about the technical infrastructure that underpins it all — it can be difficult to understand when to make bets on the early version of a technology without seeing how it will impact the world.

“In general, people are happy to throw a lot of money on gadgets,” says Dirk van der Linden, senior lecturer at the Department of Computer and Information Sciences at the UK’s Northumbria University, though he adds that even these companies aren’t thriving — finding a business model that sticks has so far eluded large parts of the animal tech space.

Like fitness and health-tracking hardware for elderly and infirm animals, which work well until animals eventually die. “People generally don’t up their subscription fees or buy new tech at that point,” says van der Linden.

Like Hirskyj-Douglas, van der Linden also stumbled into animal-computer interaction without expecting to. He got into developing tech for animals while working with a college to program a drone that would entice a dog to walk around at the University of Haifa in Israel.

“We got so interested in how ridiculous the entire process was,” he says. “There were so many challenges in trying to teach the dog to follow the drone, alongside the dog getting really angry at the drone.”

Van der Linden, who is also the co-founder of Tech4Animals, a cross-disciplinary research collective developing and studying tech that can be used by and for animals, couldn’t help but bring this interest to his academic research.

“We’re not looking to build tech for tech’s sake,” says van der Linden, “but seeing what problems we have in the real world, and how we actually put technology into the human-animal relationship to solve those problems.”

In the case of van der Linden’s lab, that includes focusing on computational animal behavior analysis, meaning that they use artificial intelligence (AI) to automate the recognition and analysis of different types of normal and pathological animal behaviors. This data helps owners figure out how best to look after their pets.

Along with his colleagues, van der Linden has been able to diagnose dogs with personality disorders that one might be able to map to those of humans, from hyperactivity and ADHD to aggression.

~

ACI is not just for pets. It can also help us better understand farm animals, including how they might feel about changes in their farmyard or home environment and how to regulate their distress. It might even help eke out more productivity from animals or ensure we spot illnesses in pets quicker.

“What’s interesting, from my perspective, is using technology to understand animals and improve their welfare, health, and so on,” says Anna Zamansky, associate professor in information systems at the University of Haifa, Israel.

Zamansky doesn’t call herself an ACI researcher — she studies animal tech and says she’s more focused on the neuroscientific and biological side of the field, which uses technology to understand how animal behavior is connected with the brain. On the other hand, ACI academics design and build things to improve interactions between animals, and between animals and humans.

“Academically, these are two different communities,” she says. “But they actually meet because they are both animal-centered.”

Zamansky also runs summer schools on “animal-centered computing.” She believes that moving closer to understanding animal psychology will open new doors for animal tech, like developing AI that can understand and parse animals' emotions and behavior — rather than building tools that gather data only by interacting with them.

“We’re not looking to build tech for tech’s sake,” says van der Linden, “but seeing what problems we have in the real world, and how we actually put technology into the human-animal relationship to solve those problems.”

“You can see really cool, futuristic research like designing a TV for dogs or a thing that the cow can use to control the temperature of its farm,” she says. “But it’s not very practical, because cows will not really use this technology in real life.”

It’d mean, for the first time, instead of making intuitive guesses about what an animal is thinking and doing, we might know for certain — we might be able to interact with these animals more deeply and take better care of them.

All of this instead of just a collar that can count the number of yards a pet has walked in a given day.

“Commercially, it’s the right thinking,” says Zamansky, referring to the collar. “But if you think about the more foundational scientific questions, we know really little about the cognition of animals.”

~

Zamansky believes that a major roadblock is the silos that ACI researchers exist in.

“The first challenge for me is the huge multidisciplinarity of the field,” she says. “You need to be computer scientists to develop AI and technology, but you can’t do it without animal experts. Usually, these two communities speak in different languages, and they almost never interact.”

Van der Linden shares a similar sentiment — except he applies it to the way the industry as a whole interacts with the outside world.

“I think it’s mostly not worked out because a lot of the field of ACI is mostly either HCI [human-computer interaction] researchers or AI researchers or veterinarians coming in,” he said. “There’s not a lot of attention from information systems people or management or business.”

Van der Linden's belief that a businessperson will help commercialize ACI projects and get them taken seriously is a common refrain in academia. But places like Y Combinator are trying to debunk that belief as a myth. Y Combinator has a guide on how to “spin your scientific research out of a university and into a startup” — their tips include working with a fellow academic to fine-tune an idea to turn into a company, internalizing that it's easy and intuitive to learn the “business” part of things, and taking the leap to leave academia and join the business world.

“There is a lot of amazing research being done by academics,” Azhynuk says, “and that research is basically waiting on the shelves to be productized. I personally cannot wait until the moment we are able to institutionally, as businesses operating in this space, put proper resources into that area.”

ACI academics want to define tomorrow — they want to ensure their creations have utility and improve the quality of life of animals, rather than weighing them down with useless tech trinkets. It’s what they’ve decreed the theme of the Ninth International Conference on Animal-Computer Interaction, to be held at van der Linden’s home institution of Northumbria University in December 2022.

More Like This