Who killed AIBO the robotic dog?

AIBO the robotic dog was hailed as a bold new plateau in robot-based entertainment on its release in 1999. Yet the device was a commercial failure. As owners and AI enthusiasts mourn its loss, it's time to ask who or what exactly screwed the pooch

Article by Michael Shea | 27 Jul 2015
  • AIBO Funeral

In 1999 Sony unveiled AIBO the robotic dog. Hailed as creating a new market in robot-based entertainment, it attracted global media attention and won plaudits for its innovative design. Yet AIBO was a commercial failure. Production had ceased by 2006 and as of 2014 year Sony no longer provides software updates or product support of any kind. As AIBO owners and enthusiasts mourn its demise, even hosting funerals for the devices as they cease to function, it is reasonable to ask who or what exactly killed AIBO.

AIBO, an acronym of Advanced Intelligence Robot (aibo 相棒 also means 'friend' or 'partner' in Japanese), operates autonomously and is fitted with sensory equipment: LCD camera, touch sensors, range finder, LED lamps, stereo mic, speakers, acceleration sensors and an angular velocity sensor. It is able to move independently, navigate and even learn behaviours it was not designed to perform. AIBO can learn up to one hundred different voice commands and is equipped with adaptive learning and growth capabilities that allow each unit to develop a unique personality with behaviour shaped by its owner. In a press statement at the time of its release, AIBO was described as being able to exhibit emotions such as “happiness and anger” as well as “the need for companionship.”

AIBO became something of a cultural phenomenon, appearing in various TV programmes including Frasier, South Park and Futurama as well as a music video for Janet Jackson. It was even satirised in The New Yorker, where it is depicted urinating nuts and bolts onto a fire hydrant. Yet in spite of such media attention and strong early sales figures – the initial batch of 3000 sold out in twenty minutes – AIBO never became commercially viable. Just six years after its launch, production was halted amid an effort to cut costs at Sony, led by Howard Stringer when he became CEO in 2005. Following this announcement, a mock funeral was staged by Toshitada Doi, AIBO’s creator, and was attended by around 100 Sony employees.

"The failure of AIBO raises various questions about what people really want from a robotic pet"

There was dismay among AIBO owners when it was announced in 2014 that product support would cease entirely. However, there are online forums where AIBO owners continue to share trivia and software updates. AIBO owners even place classified ads for 'organ donations,' whereby the parts of 'dead' AIBOs are used to repair broken devices, as spare parts are no longer being manufactured.

One possible reason AIBO failed to become a commercial success was its cost. Priced at JPY250,000 (over £1000) the product was never intended for mass market consumption. While £1000 may be significantly cheaper than the cost of keeping a live cat or dog, for a consumer entertainment product it was simply too expensive. Beyond this it is also worth questioning whether, at any price, most consumers were ready to accept the notion of a robotic pet.

In a study at Purdue University in Indiana, AIBO was placed in residential care homes to see if it would improve senior citizens’ quality of life. The principal investigator in the study, Alan Beck, describes the hostility with which many viewed the idea of replacing their pet with a robot: “I bordered on almost getting hate mail when I first started this study [from people] thinking that the study was to replace pets in people’s lives […] The truth of the matter is that people, given the choice, would prefer having the real dog.”

In another study researchers analysed AIBO online discussion forums to see how members conceptualize their robotic pet. They found that the majority viewed AIBO as having mental states and being able to form social relationships, yet very few (less than 5%) thought AIBO deserved care, rights or respect: “the relationships members had with their AIBO were remarkably one-sided. They could lavish affection on AIBO, feel companionship, and potentially garner some of the other psychological benefits of being in the company of a pet. But since the owners knew that AIBO was a technological artefact, they could ignore it whenever it was convenient or desirable to do so.”

Attempts to create robotic pets are nothing new. Sparko the Robot Dog was an exhibit at the 1939 New York World’s Fair. The animatronic dog wowed audiences with its ability to respond to light and heat. However, one night a door was left open and Sparko, attracted by the headlights of a passing car, ran out and was crushed. It would be decades before technology advanced to such an extent that mechanical pets became a reality. Japanese media art expert, Machiko Kusahara, argues that AIBO and subsequent efforts to create robot pets should be viewed as the materialisation or “incarnation” of virtual pets rather than as robotic versions of animals. The development of AIBO was arguably spurred on by the commercial success of virtual pets in the 1990s.

Tamagotchi (for anyone born after 1995) are virtual pets, usually attached to key chains that became a surprise global trend. The brand name is derived from the Japanese for ‘egg’ (tamago) and the Japanese approximation of the English ‘watch’ (uochi). The original concept was a pet that would exist inside a person’s watch and would therefore be more convenient and portable than a living animal. The product, consisting of a small egg-shaped computer with a small screen and buttons, allows the user to hatch and raise a virtual animal by attending to its frequent needs: feeding, playing etc.

Anthropologist Anne Allison analysed the success of the Tamagotchi in her book on Japanese toys, where she compares Tamagotchi to the invention of the Walkman and Karaoke, as devices that sought to make an activity more portable and accessible. She argues that the virtual pet has at least one important difference to these previous innovations, in that it generates a sense of presence in the user: “Whereas music is an experience or performance, a pet, at least as it is conventionally conceived, is a living organism – usually an animal. One of the most noted characteristics of the tamagotchi, however, and one that contributes to its popular and global appeal, is the uncanny sense of presence it generates in players.”

Why did Tamagochi succeed where AIBO failed? One answer is that virtual pets have no physical body, and therefore are unlikely to be mistaken for a live animal. The user is pleasantly surprised when they develop a feeling of presence in the device. On the other hand, AIBO, which at first glance has the appearance of a dog, is doomed to be compared with the real thing and found wanting. AIBO is like a living pet, only less intelligent, less cuddly, less responsive.

Empathy towards nonhuman entities, robotic or otherwise, is more complex than simply degrees of intelligence or life-likeness. It is quite possible to become attached to pixels on a screen and yet to be underwhelmed by AIBO, despite its genuine claim to a degree of intelligence. The failure of AIBO raises various questions about what people really want from a robotic pet.

While AIBO may have been and gone, efforts to create robots for social and entertainment purposes still continue. Last year the Japanese telecommunications company SoftBank announced its new companion robot Pepper. Equipped with face and voice recognition capabilities that allow it to respond to the user's emotional state, as demonstrated in a frankly quite terrifying promotional video, Pepper is being promoted as an emotional robot, rather than one that can provide any practical domestic assistance. “Pepper will help people grow, enhance their life, facilitate relationships, he will have fun with them, give some services and connect them with the outside world.” Pepper went on sale in Japan at the end of June.

With a price tag almost as dear as AIBO it remains to be seen whether Pepper will share its fate or if it will represent a step forward in robotic entertainment. The future of robotic pets, at least in the short term, may depend on creating devices that are not substitutes for living creatures but instead are their own distinct entities. This may be the only way to sidestep the uncanny valley and create robotic companions that people are willing to embrace.