Mar 28 2019

Robotic Pets

I warn frequently about the folly of trying to predict the future. Obviously we need to do this to some extent, but we always need to be aware of how difficult it is. It is especially hard to predict how people will use technology, even if the technology itself is inevitable. Until devices are in the hands of actual people out there in the world, who try to incorporate the tech into their daily lives, we just can’t know how it is going to shake out.

So, having said that, I am going to make a prediction about how people are going to incorporate future technology. I think robotic pets are going to be increasingly popular as the technology advances. At least I am going to build what I think is a strong case for this prediction. The risk is that there is something I cannot anticipate that will be a deal-killer. Feel free to try to shoot this down and bring up points, but hear me out first.

From a neurological point of view, I do not see any obstacle to people bonding fully with robotic pets. Neuroscience has clearly established that the human brain has certain algorithms that it uses to assign emotional significance to things, to form emotional attachments, and to respond to emotional signaling. In order for the full suite of emotional responses to be in play, being alive is simply not required. That is not how our brains work.

Our visual systems, for example, sort the world into two categories – things that have agency, and things that do not have agency. Having agency means that our brains infer that they are able to act with their own will and purpose. They infer this from how objects move. If they move in a way that cannot be explained simply as passive movement within an inertial frame of reference, then they must be moving on their own. Therefore they have agency.

Those things that appear to have agency connect to the emotional centers in our brain, and are assigned an emotional significance. We either fear them, or love them, or hunger for them, or whatever. We can trigger our emotional responses with cartoon creatures, with animals, and yes, even with robots. Our brains treat things that act alive as if they are alive – no different. Watch this video of a dancing robot to see how much we respond to movement alone.

Further, we evolved to appreciate cuteness – which is a suite of features we associate with our young. We simultaneously evolved to have certain features as children and appreciate those same features as adults. This way parents are really motivated to take care of their helpless children, which maximizes their Darwinian success. Even if the full story is more complex than this, cuteness is a thing, and things that mimic the cuteness of our children trigger a powerful emotional reaction.

In fact we domesticated pets to maximize their cuteness. Part of this process was neotony – retaining juvenile characteristics into adulthood. We bred our wolves to look more like pups.

We also evolved a very sensitive perceptual system to detect the emotions of others. These systems are so sensitive, we can detect emotional states in stick figure drawings, in emojis, in cartoons, in puppets, in animals, and yes – in robots.  We are already developing the science of social robots.

Neurosciences answer to the question of whether or not we can fully bond with robotic pets is a resounding yes. In fact, what is needed is far more minimalist than you might at first think. Consider the movie Wall-e.

One of the counter points that people often bring up when I mention this prediction is that they love the way their pets behave, not just look. This gets to AI, which has been advances at an accelerated pace. We have AI that can beat world masters in chess and go. I have no doubt we can build AI algorithms to mimic the behavior of dogs and cats, if not now then probably within a generation.

To this point, an SGU listener wrote in after discussing it on the show:

I do realize that we will find robotic “creatures” adorable and assign agency to them—I remember being outraged by the human person kicking the doglike robot (in a demonstration of how it is hard to knock over)—but my point is that I am enchanted by the fact that an animal as capricious as my cat chooses to grace me with her affection. There’s an element of earned trust that is a big part of the relationship for me. So a programmed cat would be a little like a prostitute, in that it would do what I asked/programmed it to do, but there would always be the knowledge that there was no reciprocity. Wouldn’t there?

To which I responded – Well, cats are not wild animals. They are domesticated. So in a way, we have already programmed them through breeding to give affection. They still have to bond to you specifically, but there is no reason that a robotic pet could not be programmed the same way – their behavioral algorithm will respond to people giving them attention and affection (are you old enough to remember those digital pets people had to feed and attend?). This, again, is no different than an animal.

Using the analogy of a prostitute was telling. First, our pets are already like prostitutes in that way. They trade affection for food and care, because we domesticated and bred them to be that way. But also, prostitution and stripping are highly successful industries precisely because it doesn’t matter if the behavior is “genuine” in some way. It still triggers all of our evolved emotions. This gets back to the basic concept that when our emotions are triggered it feels real, even if we know it isn’t on some level.

So I think I have made a solid argument for why we will have no problem fully emotionally bonding with a furry cute robotic pet that is programmed to push all our cuteness and nurturing buttons. But this doesn’t mean we would prefer such a pet. One potential argument against robotic pets is that they will be expensive relative to a live pet – but this is likely not to be true forever. Living pets need to be fed, and they need at least basic medical care, and they may need to be boarded at times. Also, people spend hundreds of dollars on pets, or thousands of dollars on pure bred species. Robotic pets will eventually cost less in total than biological ones.

But let’s get to the real reason people will ultimately prefer robotic pets. The possibilities for building in useful technology are endless. An advanced robotic pet could also serve as a defensive and alarm system, an alarm for smoke, carbon monoxide, and other harmful gases, a baby monitor, and even a first responder. A small CO2 canister can be triggered to blow out the mouth to put our small fires. The pet can automatically call the police or fire department, and alert you that something is going on.

These are just the most obvious things. You could recharge your cell phone off their batteries. They could be a walking cell phone. They could stream what they are seeing and hearing to an app on your phone. They would be excellent companions for older individuals, affording them more independence for longer. This in itself would be a huge cost saving – anything that keeps people out of a nursing home is cost effective.

There are also many working tasks they could perform, such as keeping vermin out of your home or away from your garden. Imagination is really the only limitation, once you have the basics of a working robot.

Further, they will be a huge advantage over living pets for the things they will not do – they will not pee or defecate in your house, shed hair, cause allergies, make your house smell, jump on your table and eat the food they find there, destroy your furniture, dig up your yard, pee on your prized plants, the occasional scratch or bite, carry infectious diseases that can transfer to people, or hump your guest’s leg. Some may argue that this unpredictable playfullness is part of the charm of having a pet (well, maybe not the disease part). You never quite know what they are going to do. Sure, I will grant that, but I don’t think it’s worth the trade-off in destruction and inconvenience.

It will also be possible to program some non-destructive unpredictable playfullness into robotic pets. You would likely also be able to choose a range of personalities. Whatever suits you.

Finally – think of all the exotic forms you could potentially have. We are not limited to cats and dogs. You want a gremlin, you got it. You could have a panther, a talking turtle, something more cartoony for the kids like elmo, a snowy owl like Harry Potter, a tiny dragon – whatever.

Once the technology crosses a certain threshold, this will be a thing. I can’t think of any deal-killer. Future generations may look at the notion of keeping a living animal in your home as primitive, disgusting, and cruel. They may have a point.

No responses yet