It’s a fun party question: which of the five types is your dog? But the study that produced those five types wasn’t built to label your dog for a quiz. It was built to fix a much grimmer number. In the United States, roughly half of the people who surrender a dog to a shelter name behavior problems as part of the reason. The researchers behind this personality work were chasing that statistic, and once you see the project through that lens, the five types stop being a personality test and start being a matchmaking tool.
That reframe matters, because the temptation with any “what type is your pet” framework is to treat the label as a destiny. The science says the opposite. Here’s what the study actually found, what each type means for the dog living in your house, and the important caveat the researchers themselves keep raising.
What the study did, in plain terms
Published in Scientific Reports, the project fed an artificial intelligence algorithm a database of around 70,000 behavioral records, then asked it to find the natural clusters, the patterns that kept showing up across breeds. The two-year effort was led by Dr. Mohammad Amirhosseini at the University of East London, alongside Professor James Serpell of the University of Pennsylvania School of Veterinary Medicine and the canine-tech company Dogvatar, which funded it.
The algorithm scored dogs on concrete, observable reactions. How did the dog respond when an unfamiliar person came into the home? Did the dog get jealous when their owner showed affection to another animal? Out of thousands of those small data points, five stable groupings emerged.
The key thing the AI added wasn’t intuition. Behaviorists have described dog temperaments for decades. What the algorithm contributed was scale: instead of one expert’s impression of a few hundred dogs, it found the seams across tens of thousands.
The matchmaking lens: which human does each dog need?
Most write-ups of this study just list the five types. That’s the paraphrase. The useful version asks a different question for each one: what kind of home does this dog actually need to thrive, and what happens when it doesn’t get it? That mismatch, the dog-home mismatch, is the thing driving the surrender statistic the researchers care about.
Excitable / hyper-attached. High energy, high attachment, low fear, a long fuse to calm down. This is the dog who follows you to the bathroom in case something exciting happens without them. “The dogs who land here aren’t badly behaved, they’re under-met,” says Dr. Mara Chen. “They need a person who can pour energy and engagement into them, and they genuinely struggle with long hours alone. Put that dog in a quiet apartment with a ten-hour workday and you’ve built a separation-anxiety case from scratch.” Right home: an active household that wants a velcro companion. Wrong home: anyone who needs a dog that self-entertains.
Anxious / fearful. High fear of both social triggers (strangers, other dogs) and non-social ones (novelty, sudden change). Cautious, deliberate, slow to commit to anything new. Interestingly, this type helps explain why two dogs with similar histories can turn out so differently. One dog can be attacked and stay friendly; another can have a minor scare and carry it for life. The wiring, not just the event, sets the depth of the response. Right home: a patient person who finds steady, low-key progress rewarding. Wrong home: a chaotic, unpredictable household.
Aloof / independent. Lower attachment and attention-seeking, higher independence, often a strong predatory streak. This is the dog who hears you call, looks at you, and trots off to finish their own business. “People read this dog as stubborn or unintelligent, and they’re neither,” Dr. Chen says. “They’re self-directed. You have to make yourself more interesting than a squirrel, which is a real engineering problem, but it’s solvable with the right rewards.” Right home: someone who respects a dog with their own agenda and enjoys the puzzle of motivating one. Wrong home: a person expecting an eager-to-please shadow.
Reactive / assertive. Heightened reactivity across situations, including toward members of their own household, and a tendency to startle hard at sudden noise, the vacuum switching on, an object hitting the floor. Many of these traits were deliberately bred into working dogs over generations; a livestock guardian that wasn’t willing to confront a threat wouldn’t have done its job. Right home: an experienced owner ready to identify triggers and build a real positive-reinforcement plan. Wrong home: a first-timer hoping the dog will “grow out of it.”
Calm / agreeable. Easygoing, quick to learn, less interested in chasing the cat or the birds. The unicorns. “These are the dogs that almost never end up in a behaviorist’s office, which is exactly why they’re underrated,” Dr. Chen notes. “They won’t win a working trial, but they slot into a family with small kids or an older owner who wants a companion that matches their pace, and they just work.” Right home: nearly any. Wrong home: a household that actively wants a high-drive working partner and finds mellow boring.
The caveat the researchers won’t let you skip
Before you decide your dog is definitively “aloof,” sit with the part the study’s own authors keep emphasizing. These categories are wide, and dogs cross between them. A single dog can be calm at home and reactive at the vet. The data also leans heavily on owner reporting, and owners describe the same behavior differently depending on the day, the setting, and their own mood. A dog labeled fearful in a noisy house might look confident in a quiet one.
So the honest use of these types is directional, not diagnostic. They’re a vocabulary for talking about tendencies, not a slot you lock your dog into. The researchers’ real hope is downstream: that shelters could one day use a tool like this to read both a dog’s tendencies and how well they’d fit a given adopter, so fewer dogs come back through the door a month later.
What newer research adds
There’s a useful tension worth knowing about. While this AI study sorts dogs by personality, a landmark 2022 study in Science from the Darwin’s Ark project, which surveyed owners of more than 18,000 dogs and sequenced DNA from over 2,000, found that breed explains only about 9 percent of the variation in a dog’s behavior. Physical traits track tightly with breed; behavior barely does.
Read together, the two findings actually agree, and they point you somewhere useful. Personality is real, substantially heritable, and worth understanding, but it lives in the individual dog in front of you, not in the breed label on their paperwork. The 2022 work did find that motor patterns, the play and movement styles inherited from hunting ancestors, are among the most heritable behaviors of all. So the practical move is to ignore the breed stereotype and watch the actual dog: how they greet a stranger, how they handle being alone, what makes them light up. That’s the personality that matters, and it’s the one a good match is built on.
References
- Amirhosseini, Mohammad, et al. “Application of Machine Learning and Deep Learning in Predicting Dog Personality Types.” Scientific Reports, 2024.
- Morrill, Kathleen, et al. “Ancestry-Inclusive Dog Genomics Challenges Popular Breed Stereotypes.” Science, vol. 376, no. 6592, 2022.
- University of East London. “AI Finds Five Dog Personality Types.” 2024.
