Why Robopets Will Never Be Real Enough

Why Robopets Will Never Be Real Enough
A boy playing with Sony's AIBO ERS-7 during AIBO's 5th anniversary exhibition held in Tokyo, Japan on May 2004. (Photo: Toshifumi Kitamura/AFP, Getty Images)

Every morning, I am stirred awake by one of the dumbest creatures in existence: a once-abandoned, now adopted 3-year-old orange tabby cat named Cheddar. In exchange for this wake up service, Cheddar gets free meals, pricy vet trips, and plenty of scritches, as do tens of millions of other pets in the U.S. alone.

The more cynical among us might say that pets are little more than expensive and far too loud roommates. Not only do you have to regularly pay attention to and feed these roomies, but oftentimes you’ll need to fork over lots of money to keep them alive, particularly for breeds predisposed to health problems. It shouldn’t be surprising, then, that for decades, some segment of the population has hoped that these furballs could one day be replaced by mechanical facsimiles with less upkeep and cost but all the benefits of domestic companionship — a robotic pet, in so many words. So, it’s worth asking: Will we be any close to that goal by 2030?

The first clear memory I have about the possibility of a real-life robopet, as I’m sure it might be for many millennials, is when the AIBO robot dog, developed by Sony, was released to the public in May 1999.

The first consumer AIBO was a lightly golden robot equipped with button sensors on each of its paws, head, and chin; LED lights that flashed green to express happiness or red for anger; and software advertised to make AIBO an “intelligent and trainable” companion that reacted to its surroundings and owner. It could also run for a whopping 1.5 hours of battery life, moved with the grace of an iceberg, and retailed for $US2,500 ($3,207). Though AIBO was certainly popular in the public imagination upon release, it never became the hottest Christmas toy out there. By 2008, Sony discontinued its annual line of AIBO dogs and ended all consumer support by 2014.

Then, like everything else these days, AIBO got a reboot. The new version, which first debuted in 2018, now retails for $US3,000 ($3,848) and boasts plenty of upgrades. For one, it just looks more dog-like than the original. It also can move much more fluidly in response to a voice command, aping the curious and unsteady shuffle of a puppy, and it even pretends to pee at inopportune times. Just looking at the two dogs side-by-side, as this YouTube video does — the differences are astonishing. It’s pretty good for a robot, all things considered, and it’s still very clearly a fake dog.

Robotics as a field is ever progressing, and there’s no denying that the next decade will bring all sorts of advancements. So the robopets of tomorrow could very well get even better at trotting and yelping like an animal. In our lifetime, we may yet have long-running robots that are nearly or fully autonomous in their ability to physically adapt to any immediate stimuli, using methods like deep learning that weave together layers of data collected from lots of training sessions that allow them to react to even unexpected situations, essentially mimicking how a brain works.

There are several research teams currently using deep learning to improve the movement capabilities of dog-like robots. And while these early efforts are geared toward creating industrial robots designed for search-and-rescue or navigation, some of these skills may trickle down to consumer robots like AIBO eventually. According to Alex Li, head of the Advanced Intelligent Robotics lab at the University of Edinburgh in the UK, it’s also possible that deep learning could one day be used to improve the social fluency of public-facing robots, bringing them closer to the sort of spontaneous yet familiar interaction that’s possible with living creatures.

“The same principle applies if you wanted to use these robots in a home environment; that robot just needs to be trained to have different behaviours — more interactive, cutesy behaviour with humans, like the way a dog sits there or moves his body so it can get a bit indulged by the owner,” Li said over the phone. It’s feasible to think that this AI training could someday allow robopets to better mimic the foibles of a pet, quirks and annoying habits alike, essentially taking the concept of the randomly peeing AIBO to a whole new level.

To be clear, though, we shouldn’t expect any consumer robot to really pull off replacing your pooch by 2030. At most, according to developmental psychiatrist Gail Melson, they could become sophisticated enough to become their own category of thing — not a living pet necessarily, but not a glorified toaster either. This might be especially true for the youngest of us growing up around them.

“What I see emerging with this kind of highly interactive technology is new categories of thought that did not exist in human thinking before,” said Melson, who has studied how people, including children, interact with current-day robots. In one study, they even asked preschool children how they felt about AIBO compared to a living dog, particularly when it came to something known as “moral standing” — the concept of an entity deserving to be treated with dignity and respect.

“I mean, they weren’t fooled. They didn’t think the AIBO was alive at all. But when we asked them about morality, they kind of hedged,” Melson said. “They didn’t think that it had the same moral standing as a living dog. But no, you couldn’t just treat it like a machine.”

Some researchers do think that robopets could become more than a passing curiosity in the near-future, but in a specific context: as companions for the lonely and/or very ill elderly. For instance, there’s already the Paro seal, advertised as an “advanced interactive therapeutic robot designed to stimulate patients with dementia, Alzheimer’s, and other cognition disorders.” Paro and others like Hasbro’s Joy for All robot cat are sometimes known as socially assistive robots. They’re outfitted with soft, snuggable fur as well as technology that allows them to cry out and respond to a person’s petting through motorised purring. They could be seen as the logical extension of doll therapy, which is sometimes used for these patients as well.

There have been studies suggesting that socially assistive robots can help relieve some of the anxiety and loneliness experienced by dementia patients or even just elderly people still able to live independently. But some of this research has also suggested that these benefits aren’t anything more than you would expect from a much cheaper plush toy (each Paro is custom-built and currently costs $US6,000 ($7,696)) or that they don’t really work for certain groups, such as those with advanced dementia.

It’s certainly possible that these robots could become a fixture in nursing homes by 2030, especially if the next generation are designed to address their current shortcomings or are packaged with other useful functions that can actively assist a elderly person’s daily living, like reminders for prescription medications. Strangely enough, though, these future robots may turn out to be more humanoid than pet-like, since some studies have suggested that people prefer the former over the latter.

But beyond their possible utility, there’s the underlying moral question of whether we should even want these robots to become “ersatz companions’’ for the elderly, according to Robert Sparrow, an Australian philosopher who first began writing about robot pets nearly two decades ago. Sparrow doubts we’ll ever come anywhere close to truly pet-like robots in our time, nor is he convinced that today’s socially assistive robots are really as beneficial as the most encouraging research suggests. But even if they were, designing these robots to do a mediocre job of pretending to be another living creature for the sick, elderly, or especially lonely is little more than deception at its best and cruelty at its worst, he argues — a trick that doesn’t really address their needs.

“It’s a failing on our part, if we think that’s a good solution,” Sparrow said. “It’s the moral equivalent of pushing a television into the room and saying, ‘Here, watch some soap operas.’ And if someone’s going to be in a room with no television, and you give them a television, no doubt that will lift their spirits a little bit. But it’s still a pretty shitty response to social isolation.”

Sparrow has long used the existence of robots and other emergent technologies as a philosophical exercise for exploring deeper issues, such as why we can form such deep bonds with living creatures like pets in the first place. Animals may not be intelligent in the same way as humans are, for instance, but in their behaviour, there’s a clear degree of otherness, something that makes them truly distinct from us. Pets also rely on us, a burden that in a way further endears them in our hearts, Sparrow notes.

“There’s this genuine need — the fact that your pets get lonely and will trash your home if you ignore them; that it will make its emotions felt. That’s part of what makes that relationship what it is,” he said. “And it’s a useful reminder, knowing that we need each other, that it’s actually part of what makes life rewarding for us. It’s in contrast to this other model, where we are sort of these atomized individuals surrounded by our technologies and our technologies that only exist to serve us, that is ultimately profoundly depressing and alienating.”

For all the nifty features and doodads that 2030’s version of the AIBO or Paro could contain, they simply won’t replicate the sheer anarchy of life that makes a pet a pet. And it’s worth wondering what would happen if these robots ever did become a household mainstay that kids or families gravitate to because they’re cheaper or don’t need to be taken to the vet, Melson asks.

“What are we losing in that kind of a scenario? I think we’re losing a lot. Because the living world is still so much richer,” she said.

Thinking back to my own cat-shaped hunk of cheese at home, I’m reminded of how utterly unpredictable he is. Sure, he’ll chirp for food or his toys several times a day and saunter around my bed most mornings, his tail lifted high in the air so that the scent of his rear pucker can waft on by. But he never makes the exact same motions from day to day, and throughout our years together, his once deeply frightful behaviour around me has melted into something approaching love, or at least into the willingness to use me as a pillow to nap on.

Cheddar may not be the brightest bulb in the drawer, but he is a furry, fleeting agent of chaos that brings a smile to my face every day. And for that gift alone, I wouldn’t trade him for all the current and future robopets in the world.