What is your pet dog really thinking? Could you attach a device to your dog and it translates what he/she is thinking? It is well known in Psychology that body signals can give an indication into human emotions, but can this work with our pet dogs? And is this accurate or, even safe? Questions like these are raised often within ACI, and more recently with commercial products for our beloved pets as, we as owners, seek further understanding: but can these products really bring this?
As a researcher within dog ACI, I don’t think you can measure continually and accurately a full scale of emotions using biometrics alone. It doesn’t account for habitual behaviour, just consensus data.
For example, while dogs tail wagging direction may be an indication (left/right) can indicate a dog’s apprehension to an approaching dog, it does not give the emotions of the dog. So when Patricia Pons asked for my thoughts about a product called DogStar TailTalk: The world’s first emotional sensor a look of horror crossed my face. They have misconstrued dog behavioral signals between approaching dogs into a fully readable biometrics device that can give a % on how happy your dog is.
This product is not alone though, with collars that text you when your dog is sad, Bowlingal an app that translates your dog barks and headsets to translate thoughts into spoken English called No More Woof.
Commercial products like these remind me of Shaun Lawson and Ben Kirmans work on fictional design within ACI in a paper for CHI ‘Problematising upstream technology through speculative design: the case of quantified cats and dogs’. They created a collar that seemed as if it could read emotions in dogs, athough the product was fake. They asked dog owners if they would trust it and found that dog owners start to trust the collar over their own vets, medical professionals and even if it was contrary to dogs displayed behaviour.
This is where the products cross over into dangerous design when you market a product with unrealistic expectations and pseudo science where the pet’s health could possibly be affected.
I think it’s like the holy grail of ACI to be able to automatically identify emotions via computer recognition but like in humans, emotions are not always easy to identify and need a multitude of methods just to suggest at emotions. Products like these do not provide accurate datasets that can be reliable. As Michelle Westerlaken said, ‘It causes people to stop looking at their dog’s behaviour, in which you can already real all those emotions and more accurately, if you have the knowledge’. While biometrics can be useful to help measure dogs emotions, they have to be used as an arsenel of tools in order to decifer accurate reading of emotions.